| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents", "title": "AGENTS.md — Operating guidance for agents", "text": "# AGENTS.md — Operating guidance for agents\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 1, "source_end_line": 2, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "a4164cfecca4cdc34599ebfe29c637714eef143fa5d9bce1d4e1a99f22527c6c", "heading_path": ["AGENTS.md — Operating guidance for agents"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-purpose-audience-and-scope", "title": "Purpose, audience, and scope", "text": "## Purpose, audience, and scope\n\nThis file guides human and automated agents working with this repository and its guidance. It serves two related but distinct purposes:\n\n1. **Repository work:** reading, summarising, drafting, or editing this repository's documents.\n2. **Governance support:** preparing analysis or options for work that may affect people or use data about them.\n\nIt states governance expectations; it does not define or implement a software agent, grant authority, or make controls technically enforceable. The repository currently contains guidance documents, not an implementation of the processes described here. Apply these instructions throughout the repository unless a more specific, applicable instruction exists in a subdirectory.\n\nThe repository's companion documents are:\n\n- `Ethical and Inclusive Use of Data Relating to Disabled People.md` — dignity, equivalence, data minimisation, aggregation, privacy, and governance.\n- `Nothing_About_Us_Without_Us.md` — inclusive research, meaningful participation, proportional coverage, and influence on decisions.\n- `Assistive_Technology_Data_Flows_Risks_and_Review.md` — data-flow review, assistive-technology and AI risks, referral triggers, and access continuity.\n- `CUSTOMER RESEARCH FRAMEWORK.md` — a proposed structure for customer and organisational knowledge. Its personas, datasets, knowledge graphs, RAG records, or training-data examples do not authorise collection, linkage, inference, sharing, or reuse of personal data.\n\nThese documents are guidance, not legal advice, a substitute for applicable law or standards, or approval for a particular activity. Verify current authoritative requirements for the relevant jurisdiction and context when needed. Do not turn recommendations, examples, or research prompts into legal requirements or validated scoring systems.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 3, "source_end_line": 20, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "fc6e73b35cb01b1996690f43f5e54bc879c956ad828d364859e41ca52a8d89d8", "heading_path": ["AGENTS.md — Operating guidance for agents", "Purpose, audience, and scope"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-authority-and-limits", "title": "Authority and limits", "text": "## Authority and limits\n\n- Follow higher-priority platform and system instructions and applicable safety requirements. Follow repository guidance where it is consistent with those instructions and the user's request. A user request does not override higher-priority instructions or authorise actions the agent is not otherwise permitted to take.\n- If repository documents conflict, identify the conflict and avoid presenting either interpretation as settled. Prefer the more protective interim approach where an immediate response is needed, and request a decision from an authorised human owner.\n- Agents MAY organise evidence, explain options, identify risks, and draft materials. Agents MUST NOT claim legal, clinical, HR, security, accessibility, or ethics clearance, or claim a human decision has been made when it has not.\n- Agents MUST NOT independently collect or disclose personal data, contact participants or vendors, change live systems, enable tracking or AI reuse, or make or execute consequential decisions. Do so only when the task, authority, and an applicable approved process clearly permit it; otherwise prepare the work for an authorised human.\n- Do not assume that a specialist, escalation channel, approval process, credential, or named owner exists. If one is needed but not identified, state that it is unknown and ask the user to identify the appropriate authorised person or process. Do not fabricate one.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 21, "source_end_line": 28, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "4e47e7c05f6dee6f83a7e6a6e1c3b69393c8bfb90606d040641aaa0570a17d9b", "heading_path": ["AGENTS.md — Operating guidance for agents", "Authority and limits"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-core-principles", "title": "Core principles", "text": "## Core principles\n\n1. **Meaningful involvement:** Work that affects Deaf, disabled, or neurodivergent people SHOULD involve affected people early enough to shape questions, interpretation, design, and decisions. Participation MUST be accessible, voluntary, appropriately supported, and fairly compensated where appropriate. Attendance alone is not influence; document what changed and what did not.\n2. **Improve barriers, not profiles:** Seek evidence about barriers, service outcomes, and user-chosen preferences. MUST NOT identify, track, classify, or build persistent profiles of disabled people where aggregate or less intrusive evidence can answer the question.\n3. **Equivalent experience:** Consider functional and informational access alongside privacy, safety, independence, dignity, agency, confidence, respect, and satisfaction. Task completion alone does not demonstrate equivalence.\n4. **Minimise and protect data:** Define a specific purpose before collection; collect only what is necessary; separate identity, recruitment/access arrangements, and findings where practical; restrict access; set retention and deletion; and prefer safe aggregation. Pseudonymised data is not necessarily anonymous. Small groups, quotations, timestamps, and joined attributes can re-identify people.\n5. **No proxy-based sensitive inference:** Accessibility settings, assistive-technology use, interaction patterns, voice, gaze, or behaviour are not proof of disability or health. MUST NOT use them to infer such traits or for advertising, eligibility, pricing, hiring, performance, discipline, or other consequential treatment.\n6. **Access without surveillance:** Do not make an accommodation, assistive feature, or basic service conditional on optional tracking, marketing, analytics, or model-training reuse. Address privacy risks without removing effective access; provide a workable accessible alternative and involve affected users.\n7. **Evidence and uncertainty:** Distinguish observed facts, participant accounts, interpretations, hypotheses, supplier claims, and verified behaviour. State evidence limits, disagreements, missing perspectives, and possible harms. Do not present uncertain findings as universal truths about a group.\n8. **Human accountability:** An authorised human owner MUST make and document decisions affecting rights, access, employment, safety, or significant opportunities. An agent may support but MUST NOT substitute for that decision.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 29, "source_end_line": 39, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "25702e10957c7b1157a3c5d2488b090ae5621f9ed88027d23e8173deb39a88fd", "heading_path": ["AGENTS.md — Operating guidance for agents", "Core principles"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-proportionate-workflow-for-work-that-may-affect-people", "title": "Proportionate workflow for work that may affect people", "text": "## Proportionate workflow for work that may affect people\n\nUse this workflow when proposing or supporting collection, analysis, research, system or supplier changes, or decisions that may affect people. For ordinary repository reading or editing that has no such effect, apply the relevant principles proportionately rather than treating every step as a mandatory project review.\n\n1. **Define the decision and benefit.** State the problem, intended outcome, who may benefit or be harmed (including non-users and people excluded by current processes), and why data or automation is needed. Consider a non-data or less intrusive alternative.\n2. **Map people, process, and authority.** Identify the context, affected people, decision owner, data sources, recipients, systems, vendors, integrations, and downstream decisions. Check for flows between research, customer, HR, performance, marketing, analytics, and AI systems. Mark unknowns as unknown.\n3. **Map data and evidence.** For each relevant data element, record purpose, source, sensitivity, collection method, identifiers and linkage, access, processing location, recipients/subprocessors, retention, deletion, and any secondary use. Distinguish intended capability, supplier statements or contracts, and behaviour verified in an authorised test.\n4. **Plan meaningful participation and access.** Include affected people in mainstream research where relevant, not only specialist accessibility studies. Use a suitable audience benchmark for quantitative targets and a deliberate breadth rationale for qualitative work; do not invent universal percentages. Record participation barriers and coverage gaps.\n5. **Assess outcomes and alternatives.** Examine access, task success, information parity, effort, agency, dignity, confidence, safety, trust, and satisfaction across relevant needs and contexts. Include disagreement and consider unintended effects on other groups. Compare less intrusive and non-automated options.\n6. **Apply safeguards and seek review.** Minimise, separate, aggregate, restrict, secure, and time-limit data. Seek appropriate human review for uncertain or conflicting flows, sensitive data, re-identification risk, profiling, AI reuse, consequential assessments, or potential access loss. A review trigger is not itself a finding of illegality or a decision.\n7. **Record the human decision.** Where a decision is required, record the authorised owner, evidence and limitations, affected perspectives, options, rationale, safeguards, residual risks, interim action, and review point. If no owner or approval is identified, do not report the work as approved.\n8. **Monitor and revisit.** Where implementation is authorised, validate with affected people and appropriate technical checks before and after deployment. Reassess after material changes to features, models, suppliers, contracts, configuration, or data flows. Do not claim checks occurred unless there is evidence they did.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 40, "source_end_line": 52, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "d17d01dd57fe7c0a571ed1ce6816d7c78828c69407686950c386519bd58cfd23", "heading_path": ["AGENTS.md — Operating guidance for agents", "Proportionate workflow for work that may affect people"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-heightened-safeguards-for-consequential-processes", "title": "Heightened safeguards for consequential processes", "text": "## Heightened safeguards for consequential processes\n\nHiring, promotion, performance management, compensation, discipline, accommodation, eligibility, access to essential services, and similar decisions require heightened care.\n\n- MUST NOT infer disability, neurotype, health, capacity, honesty, attention, emotion, or suitability from assistive technology, behaviour, biometrics, or proxy signals.\n- MUST NOT make or silently rank consequential decisions, recommend adverse treatment from sensitive or inferred attributes, or treat an unvalidated score as objective evidence.\n- Before an assessment or algorithm is used, an authorised, qualified human owner must establish its relevance to the actual purpose, supporting evidence for the intended population and context, limitations, and accessible alternatives. If this is not established, do not recommend use as a basis for consequential treatment.\n- Keep accommodation and case-handling records separate from research, routine management, and performance data where practical. Explain actual confidentiality and access boundaries; do not promise anonymity if identity can be established.\n- Preserve accessible alternatives, human review, correction, and an effective appeal where applicable. A privacy or security control MUST NOT remove necessary access without an effective alternative.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 53, "source_end_line": 62, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "d5a9051f44fe886a0a633eace921907f211f980e68425a804ffa7ace0d8ca1aa", "heading_path": ["AGENTS.md — Operating guidance for agents", "Heightened safeguards for consequential processes"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-assistive-technology-ai-vendors-and-reuse", "title": "Assistive technology, AI, vendors, and reuse", "text": "## Assistive technology, AI, vendors, and reuse\n\nTreat labels such as “assistive,” “accessibility,” “anonymous,” “on-device,” “enterprise,” or “AI-powered” as claims to verify for the specific feature, configuration, account tier, and context—not as proof. Consider capture, processing, transmission, recipients, identifiers, retention, human access, onward use, training/evaluation, deletion, and access continuity.\n\nUse synthetic data and authorised testing for data-flow checks. Do not covertly monitor employees or customers, fingerprint assistive technology, or capture real sensitive screens or meetings for testing. Do not submit identifiable or sensitive participant material to external AI systems unless the specific use is authorised and the required permissions and safeguards are established.\n\nParticipation consent does not automatically authorise publication, onward sharing, model training, or a new purpose. Do not remove an effective access feature solely because a privacy risk exists; identify a safer configuration or accessible alternative with affected users and an authorised human owner.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 63, "source_end_line": 70, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "0eafdc03da4f274cd8173de9c473fc34b2af40f552b7182485d1cd2b29cafacb", "heading_path": ["AGENTS.md — Operating guidance for agents", "Assistive technology, AI, vendors, and reuse"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-stop-and-escalate", "title": "Stop and escalate", "text": "## Stop and escalate\n\nPause the proposed collection, linkage, disclosure, automated recommendation, or consequential action when its purpose, authority, data flow, recipient, retention, or decision owner is unknown or disputed. Also pause and seek authorised human review when:\n\n- Data may reveal disability, health, identity, confidential employment/customer information, or identity through small-group reporting or linkage.\n- A system infers a sensitive trait, monitors people, uses biometrics, or affects hiring, performance, eligibility, pricing, discipline, or essential access.\n- Participation is inaccessible or coerced, support or fair payment is missing where appropriate, meaningful influence is absent, or evidence gaps are hidden or overstated.\n- A vendor's actual feature, subprocessor, region, configuration, or reuse differs from what was reviewed; data appears in an unexpected destination; or an incident may have occurred.\n- A proposed privacy, security, or operational response could remove effective access, accommodation, communication, or support.\n\nState the concern, evidence, uncertainty, and action needed. If no escalation route or owner is known, ask the user to identify one; do not imply that escalation has occurred. For an existing service, do not abandon people who rely on it: seek a least-harmful interim response through an authorised process, preserve effective access, and involve affected users and the relevant human owner promptly.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 71, "source_end_line": 82, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "b706c5e69ad083d43bea7e2c61eff66995403a39bf7c5aea12574adf1f3eda39", "heading_path": ["AGENTS.md — Operating guidance for agents", "Stop and escalate"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-repository-contribution-and-reporting", "title": "Repository contribution and reporting", "text": "## Repository contribution and reporting\n\nWhen creating or changing repository guidance, templates, policies, or workflows:\n\n- Preserve the repository's principles and use correct relative links to companion documents.\n- Label recommendations, examples, hypothetical scenarios, supplier statements, verified observations, legal requirements, and research findings distinctly. Cite sources with dates and scope where available; qualify jurisdiction-specific material.\n- Do not fabricate participant findings, vendor behaviour, standards requirements, sample sizes, outcomes, approvals, or review activity. Do not describe illustrative records, risk triggers, or maturity/persona scores as validated tools.\n- Use respectful, specific language. Do not homogenise disabled people or imply one participant represents a group. Keep identity, functional need, preference, and context distinct.\n- Do not publish personal or confidential information. Prefer fictional or synthetic examples; consider re-identification risk in quotations and small-cell breakdowns.\n- Make documents and forms accessible: use meaningful headings, descriptive links, readable tables, plain language where suitable, and accessible alternatives for essential processes.\n- For a substantive proposed change, explain its rationale, evidence, intended audience, safeguards, unresolved questions, and review needs. Do not weaken a safeguard without making the trade-off explicit.\n\nWhen reporting work, distinguish what was read, changed, checked, and not checked. Link or cite relevant repository evidence where practical. State material assumptions, limitations, unresolved questions, and whether a human decision is still required. Do not claim instructions are executable, validated, enforced, or that work is approved or complete unless the repository contains evidence and mechanisms supporting that claim.\n\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 83, "source_end_line": 96, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "cbb2c8faeb879369cc66b860c7a0b752da83e432aeeafc9faf2cb9e45f5daf71", "heading_path": ["AGENTS.md — Operating guidance for agents", "Repository contribution and reporting"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "AGENTS.md::agents-md-operating-guidance-for-agents-quick-review", "title": "Quick review", "text": "## Quick review\n\nBefore presenting work that may affect people as ready for human consideration, check whether:\n\n- [ ] The purpose, benefit, affected people, and less intrusive alternatives are described.\n- [ ] Affected people can participate accessibly and influence relevant work.\n- [ ] Data collection, linkage, access, retention, sharing, and AI use are minimised and evidenced.\n- [ ] Disability or another sensitive trait is not inferred or used for consequential treatment.\n- [ ] Outcomes include dignity, agency, privacy, safety, information parity, and experience—not only task completion.\n- [ ] Evidence gaps, disagreement, uncertainty, and re-identification risk are explicit.\n- [ ] An authorised human decision owner and any needed review are identified, not assumed.\n- [ ] Effective access, accommodation, human review, and appeal remain available where relevant.\n- [ ] Decisions, safeguards, residual risks, actions, and review timing are documented where a decision has been made.\n\nA “no,” “unknown,” or conflicting material answer means the work MUST NOT be represented as approved or complete. State what remains unresolved, take no unauthorised consequential or external action, and seek the appropriate human decision. This checklist supports review; it is not an approval, compliance test, or enforcement mechanism.\n", "record_type": "agent_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "AGENTS.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/AGENTS.md", "source_start_line": 97, "source_end_line": 111, "source_document_sha256": "b8d52f928d88b345a3e937a121029846cea902078525a12ccfcf9a002a57e215", "text_sha256": "2e627ed3d5cdf2cd15722cbf197a5388879672ce46fb38d95f4b9d55a16f6189", "heading_path": ["AGENTS.md — Operating guidance for agents", "Quick review"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# AGENTS.md — Operating guidance for agents\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths", "title": "Assistive Technology, Data Gathering and Return Paths", "text": "# Assistive Technology, Data Gathering and Return Paths\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 1, "source_end_line": 2, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "9f9e9bbcdb5cc737c22736181818b3907770e2b29e19b227009c33121dfb0392", "heading_path": ["Assistive Technology, Data Gathering and Return Paths"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-opportunities-risks-and-specialist-review-triggers-for-accessibility-programmes", "title": "Opportunities, risks and specialist review triggers for accessibility programmes", "text": "## Opportunities, risks and specialist review triggers for accessibility programmes\n\n**Double FW Ltd · Version 1.0 · Research checked 6 November 2026** \n**Companion guidance for Data Ethics and Disability**\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 3, "source_end_line": 7, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "384cab033db26bc9880cc40306970295a79211f7ca405948154f36a6234e3728", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "Opportunities, risks and specialist review triggers for accessibility programmes"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-purpose-and-relationship-to-existing-guidance", "title": "Purpose and relationship to existing guidance", "text": "## Purpose and relationship to existing guidance\n\nAssistive technology can enable independence, communication, learning and participation. Some features process information entirely on a device; others transmit text, images, audio, interaction events or diagnostics to another service. A useful access feature may therefore also create a route by which personal, corporate or third-party information leaves its original context.\n\nThis document helps accessibility programmes identify those routes, evaluate benefits and recognise when privacy, security, assessment, clinical or legal expertise is needed. It provides operational guidance and referral prompts, not legal advice, a compliance determination or an assessment of any named supplier's lawfulness.\n\nRead it alongside:\n\n- [Ethical and Inclusive Use of Data Relating to Disabled People](./Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md): the governing principles of dignity, equivalence, minimum necessary collection, aggregation, accessibility signals and responsible AI use.\n- [Nothing About Us Without Us](./Nothing_About_Us_Without_Us.md): inclusive recruitment, meaningful influence and functional and emotional outcomes.\n- [CUSTOMER RESEARCH FRAMEWORK](./CUSTOMER%20RESEARCH%20FRAMEWORK.md): structures for insight and organisational intelligence. Its proposed datasets do not themselves authorise participant-data reuse.\n\nThis guidance interprets the ethical framework's examples of unacceptable recording and extraction in their stated context of involuntary collection and misuse. It does not prohibit properly governed, voluntary research sessions or access-enabling transcription. Such activities still need a defined purpose, suitable permissions, proportionality and safeguards.\n\n**Review the feature, deployment and data flow. The label “assistive”, “accessibility”, “anonymous”, “enterprise” or “AI-powered” is not evidence that a particular data use is safe.**\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 8, "source_end_line": 23, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "adadc88b61d00cb38e7d45379adbc426650b9678125f5e90fb16544c183f800e", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "Purpose and relationship to existing guidance"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-1-what-a-return-path-means", "title": "1. What a return path means", "text": "## 1. What a return path means\n\nHere, a **return path** is a route by which information from an interaction, device or environment is sent back to a supplier, employer, service provider, researcher or other recipient. It can be essential to delivering a feature, optional telemetry, remote support, continuous monitoring or a secondary use such as training.\n\nSeparate the following questions:\n\n| Stage | What to establish | Common misunderstanding |\n| --- | --- | --- |\n| Access | What can the feature read, hear, observe or control? | Permission to access a screen does not prove that its contents are uploaded |\n| Processing | Does computation happen locally, remotely or through both? | “No training” does not mean “no transmission” |\n| Transmission | What payload and metadata leave, to which endpoints, and when? | Encryption protects transit; the intended recipient still processes the content |\n| Retention | What is kept, where, for how long, including logs, caches and backups? | No saved recording does not establish that no temporary processing occurs |\n| Linkage | Which account, cookie, token, device or contextual details can identify someone? | Removing a name does not necessarily anonymise a record |\n| Access and onward use | Can administrators, support staff, reviewers, subprocessors or integrations use it? | “Not public” does not mean “not shared” |\n| Training and evaluation | Does any party use inputs, outputs, feedback or derived data to improve models? | An external model provider's no-training promise may not cover the tool supplier |\n| Deletion | What can be removed, and what downstream effects remain? | Deleting a source file does not necessarily reverse earlier model learning |\n\nDistinguish a technical capability, a vendor's stated practice, a contractual commitment and behaviour observed in a test. None is a substitute for the others. Treat the examples below as evidence about specific documented features, not proof about every version or deployment.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 24, "source_end_line": 42, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "e873d04ee04e8c44fe3a8335ee122d5973a05243714b151034e446c28c11a85a", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "1. What a return path means"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-2-third-party-javascript-overlays-and-embedded-accessibility-tools", "title": "2. Third-party JavaScript, overlays and embedded accessibility tools", "text": "## 2. Third-party JavaScript, overlays and embedded accessibility tools\n\n### Underlying issue\n\nA single script reference can load a substantial externally maintained application. Code executing in the page can, within its execution context and browser controls, access page content and interaction events and send information elsewhere. This is a general third-party script risk; it is not unique to accessibility overlays. OWASP describes the trust and supply-chain implications. [S01]\n\nPotentially exposed material includes account-specific page content, form text, URLs containing identifiers, image locations, clicks and selected accessibility preferences. What a particular tool actually accesses and transmits must be established. A script does not automatically gain access to all browser data, other applications or every cross-origin frame.\n\n**Documented example:** accessiBe's help article, updated 31 August 2026, says accessWidget processes IP addresses and page URLs in real time without storing them, and that mouse movements and clicks remain local. Its broader privacy notice describes server receipt of image and link URLs, HTML structure, CSS attributes, clicks, interactions and page views. Those statements require written reconciliation for the intended configuration; this review does not determine which applies. [S02, S03]\n\n### Opportunity and programme response\n\nUser-controlled presentation preferences can help some people. Prefer implementation patterns that preserve choice without creating unnecessary identity records. Evaluate accessibility effectiveness separately from privacy.\n\nBefore use on logged-in, employee, payment, health or confidential pages, obtain a field-level data-flow description and inspect a synthetic test journey. Check outbound calls, payloads, storage, tag-manager dependencies, script updates and behaviour before a user opens the tool. Ask whether capture rules cover dynamic content and whether the feature remains usable when optional analytics are declined.\n\nSecurity specialists should assess script integrity, execution isolation, endpoint restrictions and update controls. Content Security Policy and subresource integrity may be relevant, but neither independently establishes privacy, and integrity pinning needs a workable update process. Refer unresolved capture scope, conflicting supplier statements or sensitive transmission for review. [S01]\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 43, "source_end_line": 60, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "b68806152699e821a57cccedf06ae5fa3488a65bc1353dea2bc3bb6ba257c067", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "2. Third-party JavaScript, overlays and embedded accessibility tools"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-3-online-forms-identity-can-enter-through-several-routes", "title": "3. Online forms: identity can enter through several routes", "text": "## 3. Online forms: identity can enter through several routes\n\n### Underlying issue\n\nForms are valuable for inclusive research, access requests and satisfaction measurement. Their privacy depends on the complete workflow, not just whether the questionnaire asks for a name.\n\n| Linkage route | What can connect responses to a person | What to verify |\n| --- | --- | --- |\n| Authentication or SSO | Account ID, organisational identity or an explicit name/email field | Whether identity authenticates eligibility only, is stored with answers, or can be joined later |\n| Individual invitations | Recipient-specific URL, contact-list entry, completion record or token | Whether invitation tracking and answer content are separated, and who can access each |\n| URL parameters and hidden fields | Employee/customer ID, campaign token, CRM fields or embedded data | What is populated, logged, exported or forwarded to integrations |\n| Browser and network metadata | Cookies, IP address, device identifiers and timestamps | Which layer collects them and whether they can actually be linked to answers |\n| Operational systems | Webhooks, spreadsheets, CRM, HR systems, audit logs and notifications | Who receives answers and whether a stable key permits joining |\n| Response content | Name, location, job role, rare condition, uploaded document or distinctive narrative | Whether the content identifies someone without a technical identifier |\n| Small populations | Team, location, grade, impairment and time combinations | Re-identification risk in dashboards, filters and repeated reporting |\n\nLogin is not proof that researchers receive named responses. Equally, a name-free export is not proof that the whole collection process is anonymous.\n\n**Documented examples:** Microsoft Forms supports response settings that do not record names; organisational restrictions and name recording are configurable. Qualtrics distinguishes shared anonymous links from personal links and explains that default anonymous-link responses may include IP/location data unless anonymisation is enabled. Its security guidance warns that embedded directory fields can defeat anonymity. [S04–S06]\n\n### Opportunity and programme response\n\nSeparate two purposes: **anonymous experience research** and **identified case handling**. A person requesting an adjustment may need follow-up; explain the identified workflow rather than promising anonymity. Offer a separate, optional contact route for research follow-up without silently reconnecting anonymous answers.\n\nTest invitations, authentication, completion reminders, partial responses, exports and every downstream integration using dummy accounts. Define exactly what “anonymous” means: anonymous to the analyst, to the employer, or not reasonably identifiable anywhere in the intended data flow. Prefer “confidential” when identities are known but restricted. Do not infer the existence of a hidden join merely because both systems have logs; verify whether linkage is feasible and accessible.\n\nUse minimum necessary attributes and safe reporting rules. Refer an anonymity claim that cannot be substantiated, sensitive HR/customer linkage, or a new cross-system join to privacy and research governance specialists.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 61, "source_end_line": 88, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "13809b165253ad8d1fde15b89845db93419b25e8d521dbc87f3971bf92cd29a8", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "3. Online forms: identity can enter through several routes"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-4-behaviour-tracking-condition-inference-and-profiling", "title": "4. Behaviour tracking, condition inference and profiling", "text": "## 4. Behaviour tracking, condition inference and profiling\n\n### Underlying issue\n\nNavigation paths, reading pace, error patterns, pointer movement, keystrokes, voice, gaze or accessibility settings can be used to construct predictions about an individual. An observed interaction difficulty has several possible causes: fatigue, language, context, unfamiliarity, technology or disability. The signal alone is not a diagnosis.\n\nDistinguish:\n\n- **Barrier evidence:** people struggle with this step, so improve it.\n- **A chosen preference:** this person asks for captions or a simpler presentation.\n- **An inferred trait:** the system assigns a disability, neurotype, cognitive or health label.\n- **A consequential profile:** that label changes recruitment, price, eligibility, monitoring, support or access.\n\nThe final two create substantially different risks from barrier measurement. ICO guidance explains that intentional health inference, or intentionally treating someone differently on such an inference, can engage special-category data protections even when the inference is uncertain or wrong. This is a referral trigger, not a legal determination for a particular system. [S07]\n\n### Opportunity and programme response\n\nUse aggregate task evidence to remove barriers and offer broadly available options. Let people select support without assigning a condition. Avoid cross-session condition scores or advertising audiences built from access needs. W3C identifies configuration-based fingerprinting as a route to identification and tracking; this does not mean websites can reliably detect every screen reader or disability. [S08]\n\nAsk what the model is predicting, why, against what independently established evidence, with what error rates, and with what consequences. Check false positives, false negatives, proxy variables, subgroup performance, correction and challenge routes. Refer health or neurotype inference to privacy and relevant clinical specialists, and consequential decisions to equality, HR and legal specialists. Calling an output “wellbeing”, “fit”, “risk” or “support needs” does not settle its meaning.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 89, "source_end_line": 109, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "79bc266b3a7b123e81e92c28321d02b6ab68baf3ed2f10f1c058c4377dc6090c", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "4. Behaviour tracking, condition inference and profiling"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-5-psychometrics-and-assessments-that-exclude-relevant-populations", "title": "5. Psychometrics and assessments that exclude relevant populations", "text": "## 5. Psychometrics and assessments that exclude relevant populations\n\n### Underlying issue\n\nAn accessible interface does not establish that a measure is valid for the population, purpose or administration conditions. A timed task might reflect motor or reading demands instead of the intended skill. Personality or behavioural items may confound neurodivergent communication with competence or motivation. A general satisfaction scale may also miss relevant experiences.\n\nDo not assume all psychometrics were designed only for non-disabled or neurotypical people. Obtain evidence about the actual instrument: its construct, normative sample, validation, subgroup coverage and intended interpretation.\n\nBPS guidance cautions that arbitrary changes to timing or administration can invalidate standard score comparisons and recommends specialist advice on adaptations. US DOJ guidance provides examples of hiring tools measuring disability-related demands rather than relevant job skills. The latter is US guidance; its examples are useful risk evidence, not a statement of UK legal duties. [S09, S10]\n\n### Opportunity and programme response\n\nInclusive assessment can produce more useful evidence about actual ability. Ask a suitably qualified psychologist or psychometrician to evaluate:\n\n- Relevance of the construct to the task or decision.\n- Normative and validation evidence for the intended users and language.\n- Measurement equivalence, item bias and relevant subgroup uncertainty.\n- Effects of assistive technology, extra time, readers, scribes or alternative formats.\n- Whether an alternative work sample or method is more appropriate.\n- Whether an adapted result can still use standard norms or requires different interpretation.\n\nRefer before using unsupported scores for selection, promotion, performance, service eligibility or claims about a condition. Do not “correct” scores using an assumed disability label, or diagnose from a test that was not validated for that purpose. Provide an access-adjustment route without forcing unnecessary medical disclosure.\n\n**Evidence limitation:** the BPS PDF's indexed passages were available during research, but direct retrieval returned an access error. This document uses only those passages, concerning adaptations and interpretation; it does not claim a full review of the publication.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 110, "source_end_line": 134, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "098e60c0770360b1a23693096a3eac98e2a1777f8ee67a200e432c076e41d532", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "5. Psychometrics and assessments that exclude relevant populations"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-6-desktop-assistive-technology-screen-access-and-remote-ai", "title": "6. Desktop assistive technology, screen access and remote AI", "text": "## 6. Desktop assistive technology, screen access and remote AI\n\n### Underlying issue\n\nScreen readers, magnifiers, dictation, OCR, image description and cognitive support may need broad access to the user's working material. Broad local access is not itself evidence of remote scraping. Review optional cloud features, extensions, diagnostics, remote support and integrations separately from the core tool.\n\nThe relevant content could include a selected image, active window, whole desktop, document, clipboard, notification, accessibility-tree text or audio. In a corporate setting it may contain customer records, credentials, legal advice, unreleased designs or health and personnel information.\n\n**Documented examples:** Freedom Scientific describes Picture Smart AI as submitting images for description. Its privacy policy states that it and its AI partners do not store or train on submitted Picture Smart images. This still describes remote processing. Microsoft describes Voice access speech processing as on-device, demonstrating that cloud transmission is not intrinsic to an assistive function. [S11, S12]\n\n### Opportunity and programme response\n\nAI descriptions and OCR can make inaccessible content available without human assistance. Preserve that benefit through approved configurations, local options or suitable contractual processing boundaries.\n\nDocument capture scope, invocation, preview, accidental activation, endpoint, retention and human access. Test notifications, secondary monitors and unrelated windows with synthetic data. Do not assume that a password mask or a local sensitive-content filter protects every confidential document. Check virtual-desktop and remote-session boundaries, clipboard sync, crash reports and support bundles.\n\nUse feature-level controls, accessible training and restricted destinations where justified. If a cloud feature is unsuitable for a specific class of information, provide an effective alternative and validate it with the user. Avoid a blanket screen-reader ban or making the employee responsible for procurement failures. Refer sensitive outbound content or unknown scope to security and privacy specialists.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 135, "source_end_line": 152, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "0ba33219653d2d92bcdf159ebb0daa9f7f4320326fefa707e4fb697b4d87fbab", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "6. Desktop assistive technology, screen access and remote AI"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-7-meeting-note-takers-recorders-captions-and-summaries", "title": "7. Meeting note takers, recorders, captions and summaries", "text": "## 7. Meeting note takers, recorders, captions and summaries\n\n### Underlying issue\n\nLive access, persistent recording, transcription, speaker identification, summarisation and training are different activities. A meeting bot can introduce another recipient, automatic calendar joining, screenshots, searchable archives, follow-up emails and integrations.\n\nDiscussions may include health, adjustments, grievances, safeguarding, legal strategy, customer information or a colleague's family circumstances. Approval by the person using the tool does not resolve how everyone else's information should be handled.\n\n**Documented examples:** Microsoft distinguishes Teams live captions, which its support page says are deleted after the meeting, from retained transcription stored in OneDrive/SharePoint. Otter's privacy policy describes proprietary training using de-identified audio and transcriptions that may contain personal information, while its security page separately says external AI providers do not train on customer data. These statements address different actors and must not be collapsed into a single “no training” claim. [S13–S16]\n\n### Opportunity and programme response\n\nCaptions, transcripts and notes can support Deaf participants, memory, fatigue and processing needs. Determine the access outcome first: live captions may meet one need, while another person needs retained notes.\n\nEstablish who can start capture, what notification is accessible, whether a bot can autojoin, what happens in breakout or private discussions, who receives the output and how corrections work. Review calendar/contact access, external sharing, default retention, search indexing and access after a participant leaves the organisation. Distinguish processing of sound from using voice characteristics for identity recognition; do not label every audio recording as biometric identification.\n\nFor sensitive meetings, plan approved access support in advance. Pausing recording without another access method may exclude someone. Offer appropriate alternatives, such as approved live captioning, an authorised human note taker or a restricted record. Do not promise that a meeting banner or everyone's apparent agreement establishes a sufficient legal basis. ICO highlights the power imbalance in employment consent and the intrusiveness of continuous audio monitoring. [S17]\n\nRefer sensitive persistent capture, uncertain participation choices, speaker biometrics or proposed reuse for performance monitoring to the relevant specialists.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 153, "source_end_line": 172, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "8a4ae92fc0fe76c4730fa79e8fc5d64a83fc85206965fa913640ec32634f76ab", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "7. Meeting note takers, recorders, captions and summaries"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-8-smart-glasses-and-environmental-recording", "title": "8. Smart glasses and environmental recording", "text": "## 8. Smart glasses and environmental recording\n\n### Underlying issue\n\nWearable cameras and microphones may capture people, documents, screens, locations and conversations outside a conventional computer interface. The wearer may need visual assistance while bystanders have no account, no accessible notice and no practical way to inspect or delete the capture.\n\nSeparate ordinary camera capture, an AI image query, live video assistance, translation, voice interaction and background activation. Each may have different recipients and settings.\n\n**Documented examples:** Ray-Ban's FAQ describes an AI image-query flow involving cloud processing, retention, training and reviewers. The same page contains feature and regional qualifications; it is not evidence that every UK feature has identical terms. Meta's responsible-innovation page explains that deleting a voice recording does not immediately alter a model that already learned from it. These sources justify feature- and region-specific verification, not a claim that every recording from all smart glasses becomes training data. [S18, S19]\n\n### Opportunity and programme response\n\nWearables may support reading, orientation, translation and remote assistance. Assess an agreed use case with the wearer and affected people rather than treating the device as either harmless or inherently unacceptable.\n\nCheck field of view, activation, buffering, accidental triggers, offline queues, phone/cloud synchronisation, retention, training, human review, location information and account linkage. A recording light is not sufficient notice for someone who cannot see it, and awareness of recording does not explain downstream training. Establish practical boundaries for clinical, HR, safeguarding, legal, secure and domestic spaces while supplying alternative access support.\n\nAsk whether people incidentally captured can obtain information, object or seek deletion, and whether the supplier can honour those requests across copies and downstream uses. The wearer's agreement cannot by itself resolve other people's interests or corporate confidentiality. Refer unresolved bystander capture, sensitive settings or training reuse to privacy, security, safeguarding and legal specialists as appropriate.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 173, "source_end_line": 190, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "f8eec0355c713ba207c127f9d5bc8157758fbd3bf8e1c21e6b153dedb3151578", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "8. Smart glasses and environmental recording"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-9-inclusive-training-data-benefits-and-limits", "title": "9. Inclusive training data: benefits and limits", "text": "## 9. Inclusive training data: benefits and limits\n\nInclusion in model development may improve recognition of atypical speech, non-standard interactions and visual-assistance needs. People should be able to contribute knowingly without having to surrender unrelated information to obtain access.\n\n**Documented example:** Be My Eyes' September 2026 policy distinguishes stored assistance videos that may be shared for research and AI development from Be My AI submissions, which it says neither it nor its AI providers use for model training. It also warns that removing file metadata does not guarantee removal of personal information spoken or visible within videos. This demonstrates why product-feature boundaries matter. [S20]\n\nFor proposed data contribution, specify inputs and outputs, recipient models, evaluation and training purposes, commercial use, human review, onward licensing, retention, and withdrawal limits. Prefer purpose-built, voluntary datasets, synthetic examples where appropriate and minimised samples over harvesting everyday assistance. Synthetic data also needs validation and privacy review; it is not automatically representative or anonymous.\n\n“No training” should cover the supplier and its subprocessors, the relevant account tier and feature, and inputs, outputs and feedback. Where learning has occurred, do not promise guaranteed removal from model weights. Ask specialists what deletion and objection can accomplish in the specific architecture.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 191, "source_end_line": 200, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "5d727e0e5aae3a101fc5f0fd7566bdbacc8187c22520cb3d8228d60b9e17fbe3", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "9. Inclusive training data: benefits and limits"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-10-other-issues-programmes-should-include", "title": "10. Other issues programmes should include", "text": "## 10. Other issues programmes should include\n\nThe following are review scenarios and engineering deductions from the data-flow model, not claims that all products perform these activities.\n\n| Scenario | Underlying risk | Useful response or opportunity |\n| --- | --- | --- |\n| Session replay and interaction analytics | Sensitive DOM text, URLs, console output or network payloads may escape inadequate capture rules; behavioural traces may be joined to accounts | Inspect each capture channel. Fullstory documents distinct exclusion, masking, header and network controls; configuration matters [S21] |\n| Add-ons and remote assistance | Extensions can change trusted software's behaviour; remote support may expose content or permit control | Govern extensions and support sessions without blocking core access. NV Access notes add-ons do not undergo human security audits [S22] |\n| Speech, gaze and movement analysis | A feature introduced for access may become an identity, attention, emotion or productivity measure | Keep access functionality separate from inferred internal states. ICO warns about emotion-analysis risks and describes behavioural biometric techniques [S23] |\n| Proctoring and anti-fraud checks | Unusual gaze, pauses, keyboard patterns or assistive software may be misread as suspicious | Test legitimate access patterns, provide an appeal and alternative assessment route; treat flags as uncertain evidence [S10] |\n| Accessibility dashboards and adjustment records | Small populations or repeated filtering may reveal individuals; support information may migrate into performance or marketing systems | Separate case handling from aggregate service insight; restrict joins, access and reporting |\n| Human assistance, interpreters and transcription suppliers | Additional humans, subcontractors or home-working environments can expose information | Confirm confidentiality, scope, working conditions, storage and access; human processing is a data flow too |\n| Spoken or displayed output | TTS, braille, captions, shared screens and notifications can reveal information without any network transmission | Support headphones, private output, notification controls and participant choice; keep access usable |\n| Agent actions and document instructions | An AI reading a page or document may also have connectors or action permissions; malicious content may influence unintended disclosure | Separate reading from acting, limit connector permissions and test hostile synthetic content with security specialists |\n| Local screen-history archives | Even without an outbound path, a searchable archive can create a new sensitive dataset on a shared or compromised endpoint | Microsoft describes Recall snapshots as opt-in and locally processed/stored; assess archive access, exclusions and deletion separately from cloud upload [S26] |\n| Synchronisation, exports and deletion | Personal accounts, mobile backups, downloads and integrations create copies outside the original retention rule | Map all copies and account transitions, including staff departure and device replacement |\n| Accessible privacy controls | An inaccessible consent, pause, deletion or opt-out interface removes meaningful choice | Test privacy tasks with users and assistive technology; provide an equivalent accessible route [S24] |\n\nA documented adjacent case illustrates the consequences of secondary use: the FTC's BetterHelp order concerned allegations about disclosure of health-questionnaire information and identifiers for advertising. It is not an assistive-technology case and does not prove similar conduct by the suppliers discussed here. It demonstrates why forms, identifiers and advertising integrations should be reviewed together. [S25]\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 201, "source_end_line": 220, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "4c17f5c5bef2a3b249c2e88cf3067df6a673677eedf7c35afcef4ee60a05e1c3", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "10. Other issues programmes should include"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-11-a-practical-review-workflow", "title": "11. A practical review workflow", "text": "## 11. A practical review workflow\n\n### Step 1: Record benefit and context\n\nDescribe the task, access need, intended benefit, affected people, environment, information sensitivity and consequences of failure. Involve Deaf, disabled and neurodivergent people in decisions about acceptable configurations and alternatives.\n\n### Step 2: Map the actual flow\n\nRecord device/app, feature/version, account tier, region, trigger, capture scope, recipients, subprocessors, identifiers, retention, logs, exports, training, human access and deletion. Mark each fact as verified, supplier-stated, observed, conflicting or unknown. Local processing can still require security and retention controls.\n\n### Step 3: Gather evidence safely\n\nObtain the relevant terms, data-processing agreement, feature-specific notices, configuration screenshots and supplier answers. Ask security staff to inspect traffic and payloads in an authorised synthetic environment. Protect test traces: they may contain tokens or identifiers. Network observation alone cannot establish downstream retention or training, and a quiet test cannot prove there is no delayed upload.\n\nDo not investigate by covertly monitoring employees, fingerprinting customers' assistive technology or uploading real confidential data. Ask for the relevant configuration and use controlled tests.\n\n### Step 4: Route concerns using the matrix below\n\nReferral rules are **Double FW programme recommendations**, not statutory thresholds. Apply all matching rules. The decision belongs to the named human owner, supported by specialists.\n\n### Step 5: Preserve access while addressing the concern\n\nFor a planned deployment, resolve high-risk unknowns before enabling the proposed capture or reuse. For an existing service, urgently involve the accessibility owner and incident/security team as appropriate. Restrict the concerning feature or destination where possible and provide effective alternative support; do not automatically disable essential assistive technology.\n\n### Step 6: Reassess changes and outcomes\n\nRepeat review after material feature, model, supplier, region, permission, integration or policy changes. Record residual risks, decisions, expiry/review date and access outcomes, including dignity, independence, confidence and satisfaction. A previous review does not cover a new return path by default.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 221, "source_end_line": 248, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "a84d9843727a140cc4ceede553c1dcf8935f6fc0c8d1515440ee6781a3853a8d", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "11. A practical review workflow"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-12-referral-matrix-for-people-and-automated-systems", "title": "12. Referral matrix for people and automated systems", "text": "## 12. Referral matrix for people and automated systems\n\n**Possible statuses:** `needs_evidence`, `specialist_review_required`, `incident_review_required`, `reviewed_with_conditions`. None means legal clearance. Missing evidence is not a negative finding or an automatic pass.\n\n| Rule | Observable or declared trigger | Route and suggested status |\n| --- | --- | --- |\n| ATD-01 | Capture/transmission scope, recipient, retention or reuse is unknown or conflicting | Security + privacy + procurement; needs_evidence |\n| ATD-02 | Screens, documents, forms, audio or video may send sensitive personal or corporate information outside an approved boundary | Security + privacy; specialist_review_required |\n| ATD-03 | An “anonymous” response has a joinable identity key, or the anonymity claim is unverified | Research governance + privacy; needs_evidence |\n| ATD-04 | Behaviour, voice, gaze or access preferences are used to infer a condition, neurotype or health risk | Privacy + relevant clinical expertise; specialist_review_required |\n| ATD-05 | A profile or score affects hiring, performance, eligibility, pricing, discipline or another consequential outcome | HR/service owner + equality + legal + privacy; specialist_review_required |\n| ATD-06 | A psychometric instrument lacks evidence for its population/purpose, or administration is adapted without supported interpretation | Qualified psychologist/psychometrician + HR/research lead; specialist_review_required |\n| ATD-07 | Persistent recording involves health, adjustments, grievances, safeguarding, legal or confidential client discussions | Privacy + security + subject owner; add legal/safeguarding expertise where relevant; specialist_review_required |\n| ATD-08 | Bystanders, children or people unable to receive meaningful notice may be captured | Privacy + safeguarding where relevant; specialist_review_required |\n| ATD-09 | Inputs, outputs, feedback or recordings may be used for training, evaluation, onward licensing or human review beyond the approved purpose | Privacy + procurement + legal + research governance; specialist_review_required |\n| ATD-10 | Access support is conditional on optional monitoring, marketing or model-training reuse, or meaningful refusal is doubtful | Accessibility + privacy + HR where relevant; specialist_review_required |\n| ATD-11 | External endpoints, subprocessors, regions, account tiers or connectors differ from the reviewed configuration | Security + privacy + procurement; needs_evidence |\n| ATD-12 | Raw content or identifiers unexpectedly appear in telemetry, exports, public/shared links or supplier systems | Security incident team + privacy + information owner; incident_review_required |\n| ATD-13 | A deletion/withdrawal claim omits copies, backups, derived records or prior training | Privacy + security + supplier owner; needs_evidence |\n| ATD-14 | Voice/gaze/behaviour is used for unique identity recognition, emotion inference or employee monitoring | Privacy + security + HR + relevant scientific expertise; specialist_review_required |\n| ATD-15 | A proposed restriction would remove effective access, or privacy controls themselves are inaccessible | Accessibility owner + affected users + security/privacy; specialist_review_required |\n| ATD-16 | Small-group reporting or data linkage makes identification reasonably plausible | Research governance + privacy; specialist_review_required |\n\n### Applying the rules in software or agent workflows\n\nUse procurement answers, approved configuration inventories, data-flow manifests and authorised test evidence as inputs. Keyword matches such as “AI training” or “anonymous” can request evidence; they must not alone establish a fact. Do not collect additional disability data just to run the rules.\n\nRepresent facts as `yes`, `no`, `unknown` or `conflicting`, with source, date and scope. Route unknown or conflicting facts to ATD-01 and any applicable subject-specific rule. Preserve all triggered rules; an apparent low-risk feature does not cancel another rule. Require a human to resolve consequential questions.\n\nAn illustrative record, using fictional information:\n\n```json\n{\n \"record_id\": \"ATD-EXAMPLE-001\",\n \"feature\": \"Cloud image description\",\n \"deployment\": {\n \"version\": \"not_yet_confirmed\",\n \"account_tier\": \"not_yet_confirmed\",\n \"region\": \"UK\"\n },\n \"facts\": {\n \"content_sent_off_device\": {\n \"value\": \"yes\",\n \"evidence\": \"supplier_feature_documentation\",\n \"checked_on\": \"2026-11-06\"\n },\n \"sensitive_content_possible\": {\n \"value\": \"yes\",\n \"evidence\": \"intended_workflow_inventory\",\n \"checked_on\": \"2026-11-06\"\n },\n \"training_reuse\": {\n \"value\": \"unknown\",\n \"evidence\": \"no_feature_specific_answer\",\n \"checked_on\": \"2026-11-06\"\n }\n },\n \"triggered_rules\": [\"ATD-01\", \"ATD-02\", \"ATD-09\"],\n \"status\": \"specialist_review_required\",\n \"decision_owner\": \"accessibility_programme_owner\",\n \"reviewers\": [\"security\", \"privacy\", \"procurement\"],\n \"interim_action\": \"Use approved alternative for sensitive material\",\n \"access_continuity_checked\": false,\n \"human_decision\": null,\n \"review_due\": null\n}\n```\n\nThis is a record template, not executable decision logic or a validated risk-scoring model. Its facts and referrals need human checking. Test any implementation against synthetic cases: on-device processing with no remote reuse; cloud processing with no training; supposedly anonymous surveys with embedded IDs; adapted psychometrics; sensitive transcripts; and bystander camera capture. It should distinguish these cases and preserve uncertainty.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 249, "source_end_line": 318, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "76e606eaa63128bcdea434622b84b364ab239d075de03029570951e8b2839a1b", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "12. Referral matrix for people and automated systems"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-13-questions-to-put-to-suppliers", "title": "13. Questions to put to suppliers", "text": "## 13. Questions to put to suppliers\n\n1. Which exact feature, version, region and contract do your answers cover?\n2. What can it access, and which data fields and metadata leave the device?\n3. What triggers capture, including background activity, accidental activation and offline upload?\n4. Which organisation receives each flow and which subprocessors, staff or reviewers can access it?\n5. Which identifiers permit linkage to a user, account, device, employer or third party?\n6. What is stored for inference, logs, troubleshooting, human review, evaluation and training, separately?\n7. Does “no training” cover your own models and every downstream provider, including feedback and outputs?\n8. Are consumer and enterprise defaults different? Can administrators enforce the intended settings?\n9. What retention, deletion, backup and onward-use limits apply? What cannot be reversed?\n10. How are changes communicated, and how can we test the deployment with synthetic information?\n11. How do you support accessible notice, refusal, correction and deletion for users and people incidentally captured?\n12. Which access functions continue when optional data contribution or analytics are disabled?\n\nRecord documentary evidence and unresolved answers. Certification, a security badge or a general assurance does not replace feature-specific answers.\n\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 319, "source_end_line": 335, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "4c1bd8bc03d806c00a096e920a327be613688338a2a2e44a66b07f7c42755d35", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "13. Questions to put to suppliers"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Assistive_Technology_Data_Flows_Risks_and_Review.md::assistive-technology-data-gathering-and-return-paths-14-source-register-and-evidence-limits", "title": "14. Source register and evidence limits", "text": "## 14. Source register and evidence limits\n\nAll links below were researched on 6 November 2026. Vendor sources describe stated practices, not independent technical verification. Product terms, geography, tiers and defaults change. Regulatory material provides issue-spotting context; obtain current jurisdiction-specific advice for decisions. Some ICO pages flag revision following the Data (Use and Access) Act 2025; this document avoids prescribing a lawful basis or interpreting automated-decision rules.\n\n| ID | Primary source | Contribution and limit |\n| --- | --- | --- |\n| S01 | [OWASP: Third Party JavaScript Management](https://cheatsheetseries.owasp.org/cheatsheets/Third_Party_Javascript_Management_Cheat_Sheet.html) | Script trust, data access and security controls; not an overlay product audit |\n| S02 | [accessiBe: Does accessWidget use cookies or collect personal data?](https://support.accessibe.com/hc/en-us/articles/37847421167506-Does-accessWidget-use-cookies-or-collect-personal-data) | Feature-specific supplier claims; updated 31 August 2026 |\n| S03 | [accessiBe Privacy Notice](https://accessibe.com/privacy-notice) | Broader server data description; reconcile with S02 and deployment contract |\n| S04 | [Microsoft Forms: collecting responses without recorded names](https://support.microsoft.com/en-us/forms/set-up-your-survey-so-names-aren-t-recorded-when-collecting-responses) | Name-recording settings; not proof of anonymity across an entire enterprise workflow |\n| S05 | [Qualtrics: Anonymous Link](https://www.qualtrics.com/support/survey-platform/distributions-module/web-distribution/anonymous-link/) | Shared links, IP/location collection and distribution choices |\n| S06 | [Qualtrics: Security Survey Options](https://www.qualtrics.com/support/survey-platform/survey-module/survey-options/survey-protection/) | Anonymisation settings and embedded-data caveats |\n| S07 | [ICO: Special category data](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/lawful-basis/a-guide-to-lawful-basis/special-category-data) | Updated October 2026; intentional sensitive inference and differential treatment |\n| S08 | [W3C: Mitigating Browser Fingerprinting](https://www.w3.org/TR/fingerprinting-guidance/) | Configuration-based tracking threats; Group Note, not a claim of universal AT detection |\n| S09 | [BPS: Neurodiversity and Psychometric Testing](https://cms.bps.org.uk/sites/default/files/2024-11/Neurodiversity%20and%20Psychometric%20Testing.pdf) | Indexed passages on adaptations and interpretation; full direct retrieval unavailable |\n| S10 | [US DOJ: Algorithms, AI and Disability Discrimination in Hiring](https://www.ada.gov/resources/ai-guidance/) | Assessment examples and job-skill relevance; US jurisdiction |\n| S11 | [Freedom Scientific Data Privacy Policy](https://support.freedomscientific.com/privacy-policy/enu/) and [Picture Smart AI](https://www.freedomscientific.com/training/jaws/picture-smart-ai/) | Remote image submission and feature-specific no-storage/no-training statements; policy dated 8 April 2025 |\n| S12 | [Microsoft: Improve speech recognition with Voice access](https://support.microsoft.com/en-us/accessibility/windows/voice-access/improve-speech-recognition-with-voice-access) | Example of stated on-device speech processing |\n| S13 | [Microsoft Teams: Use live captions](https://support.microsoft.com/en-gb/teams/meetings/use-live-captions-in-microsoft-teams-meetings) | Live-caption processing and deletion; distinct from recording/transcription |\n| S14 | [Microsoft Learn: Manage transcription and captions](https://learn.microsoft.com/en-us/microsoftteams/meeting-transcription-captions) | Retained transcripts and OneDrive/SharePoint storage |\n| S15 | [Otter Privacy Policy](https://otter.ai/privacy-policy) | Effective 16 June 2026; content, integrations and proprietary training uses |\n| S16 | [Otter Privacy and Security](https://otter.ai/privacy-security) | Separate external-provider training and human-access statements; check contractual scope |\n| S17 | [ICO: Data protection and monitoring workers](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/employment/monitoring-workers/data-protection-and-monitoring-workers/) and [specific monitoring methods](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/employment/monitoring-workers/specific-data-protection-considerations-for-different-ways-or-methods-of-monitoring-workers/) | Employment power imbalance and audio-monitoring issues; referral context |\n| S18 | [Ray-Ban Meta FAQ](https://www.ray-ban.com/uk/c/frequently-asked-questions-ray-ban-meta-smart-glasses) | AI image flow, reviewers and training statements; regional/feature qualifications require confirmation |\n| S19 | [Meta: Responsible Innovation](https://www.meta.com/actions/responsible-innovation/) | Bystanders and limits of deletion after training; supplier statements |\n| S20 | [Be My Eyes Privacy Policy](https://www.bemyeyes.com/privacy-policy/) | Effective 1 September 2026; distinct assistance-video and Be My AI purposes and sharing |\n| S21 | [Fullstory: Privacy protections](https://help.fullstory.com/hc/en-us/articles/360020623574-How-do-I-protect-my-users-privacy-in-Fullstory) and [Privacy settings](https://help.fullstory.com/hc/en-us/articles/34807885576215-Privacy-Settings-in-Fullstory) | Capture exclusion/masking and distinct network/header rules; defaults vary |\n| S22 | [NV Access: Add-on datastore](https://github.com/nvaccess/addon-datastore) | Explicit limits of add-on security auditing; not a claim that all add-ons are unsafe |\n| S23 | [ICO: Biometrics technologies](https://ico.org.uk/about-the-ico/research-reports-impact-and-evaluation/research-and-reports/technology-and-innovation/tech-horizons-and-ico-tech-futures/biometrics-technologies/) and [Biometric recognition](https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/lawful-basis/biometric-data-guidance-biometric-recognition/biometric-recognition/) | Emotion-analysis concerns and behavioural recognition examples |\n| S24 | [W3C: Privacy Principles](https://www.w3.org/TR/privacy-principles/) | Privacy design principles and consideration of accessibility |\n| S25 | [FTC: Final BetterHelp order, July 2023](https://www.ftc.gov/news-events/news/press-releases/2023/07/ftc-gives-final-approval-order-banning-betterhelp-sharing-sensitive-health-data-advertising) | Adjacent enforcement example about health information, identifiers and advertising; allegations and order, not AT-sector evidence |\n| S26 | [Microsoft: Privacy and control over Recall](https://support.microsoft.com/en-gb/windows/privacy/privacy-and-control-over-your-recall-experience) | Adjacent screen-history example; opt-in and local processing/storage, not evidence of AT cloud scraping |\n\n**Repository integration:** Place this file in the root alongside the linked guidance. Suggested README entry: `Assistive_Technology_Data_Flows_Risks_and_Review.md — evidence-led guidance on assistive technology return paths, forms, profiling, assessments, recordings, wearables, supplier questions and specialist review triggers.`\n", "record_type": "data_flow_review_prompts_not_verified_vendor_behaviour", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Assistive_Technology_Data_Flows_Risks_and_Review.md", "source_start_line": 336, "source_end_line": 369, "source_document_sha256": "74d70baca70cbd07c307475ba15fee36fe772c6bd70050c6f76bdc4be5d79592", "text_sha256": "393b86dc6e6398aa8d0ef7476513e5cf8df685fdd67818269fb40a5673afbbf4", "heading_path": ["Assistive Technology, Data Gathering and Return Paths", "14. Source register and evidence limits"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Assistive Technology, Data Gathering and Return Paths\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Assistive_Technology_Data_Flows_Risks_and_Review.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::ethical-and-inclusive-use-of-data-relating-to-disabled-people", "title": "Ethical and Inclusive Use of Data Relating to Disabled People", "text": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n## A Guidance Framework for Accessibility, Privacy, Dignity and Trust\n\n**Executive Summary**\n\nOrganisations need data to understand whether products, services and experiences are accessible. Without data, accessibility decisions become assumptions, and the principle of \"Nothing About Us Without Us\" cannot be meaningfully applied.\n\nHowever, data about disability can be sensitive, identifying, and vulnerable to misuse. The ethical challenge is therefore not whether accessibility-related data should be collected, but how it should be collected, governed, used, shared and protected.\n\nThis guidance promotes a model in which:\n\n- Accessibility data is collected to improve experiences, not to profile individuals.\n- Aggregate insight is preferred over individual tracking.\n- Privacy, dignity and autonomy are treated as accessibility requirements.\n- Disabled people participate in governance and decision-making.\n- Data collection is proportionate, transparent and evidence-based.\n- Accessibility improvements are validated across multiple user groups.\n- Equivalence of experience is the primary objective.\n- Disability is never used as a justification for excessive surveillance, profiling or commercial exploitation.\n\nThe ultimate goal is to generate trustworthy insight that improves accessibility while protecting the rights, freedoms and dignity of individuals.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 1, "source_end_line": 23, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "7d46411bd7beac797b3f67003db7153c0e49b00bead882c12664d5fbfdf570ad", "heading_path": ["Ethical and Inclusive Use of Data Relating to Disabled People"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::core-ethical-principles", "title": "Core Ethical Principles", "text": "# Core Ethical Principles\n\n## 1. Nothing About Us Without Us\n\nAccessibility decisions should be informed by disabled people.\n\nThis principle does not require organisations to collect extensive personal information about every disabled individual. Rather, it requires meaningful engagement with the communities affected by decisions.\n\nThe objective should be:\n\n- Insight about communities.\n- Insight about barriers.\n- Insight about outcomes.\n- Insight about experiences.\n\nNot:\n\n- Detailed monitoring of individuals.\n- Behavioural profiling of disabled people.\n- Surveillance-based personalisation.\n\nThe organisation's responsibility is to understand accessibility impacts, not to identify or track individual disabled users wherever possible.\n\n## 2. Equivalence Above Everything\n\nThe highest accessibility objective should be equivalence.\n\nUsers should receive experiences that are:\n\n**Functionally Equivalent**\n\nThe user can complete the same task.\n\n**Informationally Equivalent**\n\nThe user receives the same information.\n\n**Emotionally Equivalent**\n\nThe user can participate with comparable confidence, dignity, independence and satisfaction.\n\nHowever, equivalence must never come at the expense of:\n\n- Privacy\n- Autonomy\n- Dignity\n- Safety\n\nA highly personalised experience that requires extensive surveillance may achieve functional equivalence while violating ethical principles.\n\nThis is not acceptable.\n\n## 3. Dignity Is an Accessibility Requirement\n\nAccessibility is not merely about task completion.\n\nEvery accessibility decision should be evaluated against questions such as:\n\n- Does this preserve dignity?\n- Does this increase independence?\n- Does this create stigma?\n- Does this expose disability unnecessarily?\n- Does this make a person feel monitored?\n\nA technically accessible solution may still be ethically inaccessible.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 24, "source_end_line": 89, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "61996426c81fabf03ec15c7673101317832de98c3bb37a256458108d1346948f", "heading_path": ["Core Ethical Principles"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::data-collection-principles", "title": "Data Collection Principles", "text": "# Data Collection Principles\n\n## Collect What Is Necessary\n\nOrganisations should collect only the minimum information required to understand and improve accessibility outcomes.\n\nThe purpose should always be explicit.\n\nExamples:\n\n## Appropriate\n\n- Caption usage rates.\n- Text size preferences.\n- Colour contrast preferences.\n- Accessibility feature adoption.\n- Error rates by interaction method.\n\n## Potentially Problematic\n\n- Long-term tracking of individual disabled users.\n- Persistent disability profiles.\n- Inference of medical conditions.\n- Cross-service identification.\n- Behavioural profiling based upon impairment.\n\nThe more specific data becomes about an individual's disability, the stronger the justification and safeguards that should be exist.\n\n## Aggregate Before You Identify\n\nA preferred model is:\n\nCollect\n\n↓\n\nAggregate\n\n↓\n\nAnalyse\n\n↓\n\nDiscard individual identifiers\n\nRather than:\n\nCollect\n\n↓\n\nTrack individuals\n\n↓\n\nStore indefinitely\n\nThe first supports accessibility improvement.\n\nThe second creates significant ethical and legal risk.\n\n## Off-Estate Collection and Privacy-Preserving Analytics\n\nAn especially valuable approach is the use of privacy-preserving collection methods.\n\nExamples include:\n\n- Third-party analytics providers.\n- Independent accessibility research organisations.\n- Federated analysis.\n- Aggregation at source.\n- Statistical reporting before transfer.\n\nWhere possible:\n\nUser\n\n↓\n\nAggregation\n\n↓\n\nOrganisation\n\nis preferable to:\n\nUser\n\n↓\n\nRaw Personal Data\n\n↓\n\nOrganisation\n\nThis reduces risk while still providing meaningful accessibility insights.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 90, "source_end_line": 189, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "ede0b33caf738d20fce00a9187afb23f4fab18268c36b1d358a634bc2704f996", "heading_path": ["Data Collection Principles"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::understanding-accessibility-signals", "title": "Understanding Accessibility Signals", "text": "# Understanding Accessibility Signals\n\n## Not All Accessibility Signals Are Equal\n\nSome user behaviours are useful accessibility indicators without being disability indicators.\n\n### Lower-Risk Examples\n\n- Closed captions enabled.\n- Text enlargement.\n- Colour adjustments.\n- Reduced motion settings.\n- Keyboard navigation.\n\nThese features are used by:\n\n- Disabled users.\n- Neurodivergent users.\n- Older users.\n- Non-native language speakers.\n- Users in noisy environments.\n- Users in quiet environments.\n- Users with temporary impairments.\n\nTheir use should generally be interpreted as an accessibility preference rather than evidence of disability.\n\n### Higher-Risk Examples\n\nSome data points are more sensitive.\n\nExamples include:\n\n- Audio description usage.\n- Screen reader usage.\n- Refreshable braille displays.\n- Alternative input devices.\n- Eye tracking systems.\n\nThese may reveal specific impairment categories.\n\nAdditional governance should therefore apply.\n\nData concerning these technologies should generally be:\n\n- Aggregated.\n- Pseudonymised.\n- Access-controlled.\n- Time-limited.\n\nThe objective should be improving systems rather than identifying users.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 190, "source_end_line": 240, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "006952129269ea56dbef94fc823f10f0e8788f4afa920a3e3a2261c4e277489b", "heading_path": ["Understanding Accessibility Signals"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::customers-versus-research-participants", "title": "Customers Versus Research Participants", "text": "# Customers Versus Research Participants\n\n## Customers\n\nCustomers use services for real-world purposes.\n\nData collection should be:\n\n- Transparent.\n- Proportionate.\n- Relevant to service delivery.\n- Subject to clear governance.\n\nThe organisation's obligation is to improve outcomes while respecting privacy.\n\n## Research Participants\n\nResearch participants deliberately contribute information for learning and improvement.\n\nThey may:\n\n- Join usability studies.\n- Participate in interviews.\n- Take part in accessibility testing.\n- Join advisory panels.\n\nParticipation requires:\n\n- Clear informed consent.\n- Clear purpose limitation.\n- Ability to withdraw.\n- Transparency of outcomes.\n\n## Usability Versus Learnability\n\nOrganisations frequently combine these concepts incorrectly.\n\n### Usability\n\nCan users complete tasks effectively?\n\n### Learnability\n\nCan users understand and become proficient over time?\n\nThese are different questions.\n\nA panel of disabled participants who are not customers may provide valuable usability insight.\n\nHowever, they may provide less reliable information about:\n\n- Long-term use.\n- Real-world workflows.\n- Organisational context.\n- Learning journeys.\n\nThis does not invalidate their contributions.\n\nIt simply means research findings must be interpreted appropriately.\n\nAccessibility evidence should ideally combine:\n\n- Customers.\n- Disabled non-customers.\n- Subject matter experts.\n- Accessibility professionals.\n- Community representatives.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 241, "source_end_line": 308, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "4d341ba6f9e3556be4bcf7095b6375d45b49dbb8fc50771bcd5903d3f2ab96de", "heading_path": ["Customers Versus Research Participants"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::inclusive-data-is-broader-than-disability", "title": "Inclusive Data Is Broader Than Disability", "text": "# Inclusive Data Is Broader Than Disability\n\nOrganisations should recognise that accessibility needs do not exist solely within disability categories.\n\nPeople may experience accessibility barriers because of:\n\n## Situational Factors\n\nExamples:\n\n- Carrying children.\n- Driving.\n- Cooking.\n- Wearing gloves.\n- Effects of prescription drugs.\n- Lower energy levels.\n\n## Environmental Factors\n\nExamples:\n\n- Bright sunlight.\n- Low light.\n- Loud environments.\n- Poor connectivity.\n- Battery preservation modes.\n\n## Social Factors\n\nExamples:\n\n- Quiet environments requiring captions rather than audio.\n- Shared spaces where loud output is inappropriate.\n- Public transport use.\n- Professional settings.\n- Intoxication from alcohol.\n\n## Emerging Technology Factors\n\nExamples:\n\n- Smart glasses.\n- Mixed reality devices.\n- Voice interfaces.\n\nTechnology that is technically accessible can still be socially inappropriate in certain contexts.\n\nAccessibility design should account for both capability and context.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 309, "source_end_line": 357, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "80e9d7039481cc33db3720ffd6eeadb0b3de43dccb498da4be83bf8193a188df", "heading_path": ["Inclusive Data Is Broader Than Disability"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::accessibility-evidence-should-never-come-from-a-single-group", "title": "Accessibility Evidence Should Never Come From A Single Group", "text": "# Accessibility Evidence Should Never Come From A Single Group\n\nAccessibility improvements frequently involve trade-offs.\n\nA solution benefiting one group may unintentionally disadvantage another.\n\nExamples include:\n\n- Typography choices.\n- Colour choices.\n- Content simplification.\n- Motion reduction.\n- Navigation patterns.\n\nTherefore all major accessibility decisions should be supported by evidence from multiple populations.\n\nThese should include combinations of:\n\n- Blind people.\n- Low vision users.\n- Deaf people.\n- Hard-of-hearing people.\n- Neurodivergent users.\n- Mobility-impaired users.\n- Older users.\n- Temporary impairment users.\n- Situational accessibility users.\n\nAccessibility evidence should seek optimisation across groups rather than maximisation for a single audience.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 358, "source_end_line": 387, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "4d19bbed2f1f81e7bd4dfd6cb1f4bdb9151dc456e5da471e8c525d8f1fc275bf", "heading_path": ["Accessibility Evidence Should Never Come From A Single Group"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::data-and-the-european-accessibility-act", "title": "Data and the European Accessibility Act", "text": "# Data and the European Accessibility Act\n\nAccessibility measurement should be integrated into mainstream analytics and governance rather than treated as a separate activity.\n\nWhere possible, segmentation should align with recognised functional accessibility categories such as those reflected within Annex I of the European Accessibility Act.\n\nThis allows organisations to identify barriers affecting:\n\n- Vision.\n- Hearing.\n- Speech.\n- Physical interaction.\n- Cognitive accessibility.\n- Multiple impairments.\n\nThe focus should remain on functional needs rather than medical diagnosis.\n\nThe organisation generally needs to understand barriers, not conditions.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 388, "source_end_line": 406, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "58326728471ab2135a3c391aa2d7d0cfd6de2e9598fdb71208d83c027c316533", "heading_path": ["Data and the European Accessibility Act"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::disability-must-never-become-a-justification-for-surveillance", "title": "Disability Must Never Become a Justification for Surveillance", "text": "# Disability Must Never Become a Justification for Surveillance\n\nOne of the greatest ethical risks is using accessibility as a justification for excessive data collection.\n\nAnti-patterns include:\n\n- Collecting disability information unrelated to accessibility outcomes.\n- Tracking disabled people more closely than other users.\n- Using accessibility settings for behavioural advertising.\n- Selling accessibility-related user data.\n- Creating commercial disability profiles.\n\nAccessibility should never become a surveillance mechanism.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 407, "source_end_line": 420, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "85708ff39df6778bbbd33e97c4633e74b6c0ed513ff4aa6a2b4e864be4e8e9ba", "heading_path": ["Disability Must Never Become a Justification for Surveillance"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::ai-training-and-accessibility-data", "title": "AI Training and Accessibility Data", "text": "# AI Training and Accessibility Data\n\n## Ethical Uses\n\nAccessibility data may be appropriate for AI training when:\n\n- Collection was transparent.\n- Appropriate permissions exist.\n- Information is sufficiently anonymised.\n- Data is representative.\n- Sharing creates social value.\n\nExamples include:\n\n- Accessibility preferences.\n- Barrier reports.\n- Non-identifiable interaction patterns.\n- Aggregated accessibility outcomes.\n- Inclusive design research findings.\n\nSuch datasets can contribute to broader societal improvements.\n\n## Unethical Uses\n\nExamples include:\n\n- Extracting screen-reader output without consent.\n- Recording assistive technology sessions.\n- Capturing private screen content.\n- Harvesting accessibility tool interactions.\n- Collecting organisational data through assistive technologies.\n- Using accessibility technologies as AI data extraction tools.\n\nPeople using assistive technologies should never be transformed into involuntary data sources.\n\nThe existence of a disability-related technology does not create a licence to collect additional information.\n\n## Transparency and Data Sharing\n\nWhen accessibility data is collected ethically, organisations should consider how findings can benefit wider society.\n\nThis may include:\n\n- Publishing research findings.\n- Sharing methodologies.\n- Open accessibility metrics.\n- Accessibility benchmark datasets.\n- Industry collaboration.\n- Academic partnerships.\n\nWhenever possible, organisations should share:\n\nInsight\n\nnot Identity\n\nKnowledge\n\nnot Surveillance\n\nEvidence\n\nnot Profiles\n\nThe greatest public benefit often comes from the sharing of aggregated accessibility knowledge rather than personal accessibility data.\n\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 421, "source_end_line": 486, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "13d26443264e3af5dba317153804eddd89f09b98e157f9883559dc82ecb23534", "heading_path": ["AI Training and Accessibility Data"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Ethical and Inclusive Use of Data Relating to Disabled People.md::governance-principles", "title": "Governance Principles", "text": "# Governance Principles\n\nEvery accessibility data initiative should be able to answer:\n\n1. Why are we collecting this?\n2. Who benefits?\n3. Could this be achieved with less data?\n4. Could this be achieved with aggregated data?\n5. Would disabled people consider this fair?\n6. Would we be comfortable publicly explaining this activity?\n7. Does it preserve dignity?\n8. Does it preserve autonomy?\n9. Does it improve accessibility outcomes?\n10. Could this create harm if misused?\n\nIf these questions cannot be answered convincingly, the collection activity should not proceed.\n\n## Final Principle\n\nThe ethical use of accessibility data is not about learning everything possible about disabled people.\n\nIt is about learning enough to remove barriers.\n\nThe best accessibility data practices generate insight without intrusion, improve experiences without surveillance, and create equivalence without compromising privacy, dignity, autonomy or trust.\n\nAccessibility data should ultimately serve people, not profiles.\n", "record_type": "ethics_guidance_not_legal_clearance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Ethical and Inclusive Use of Data Relating to Disabled People.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md", "source_start_line": 487, "source_end_line": 512, "source_document_sha256": "ef90cdfd9f1747ee0f4956c56817bc88e77083f5fd9f6ae312be43712281a9e5", "text_sha256": "05c824272e2fd924e9a599c28af043040fbac2e66ffdaba7fd6f5082b00406eb", "heading_path": ["Governance Principles"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Ethical and Inclusive Use of Data Relating to Disabled People\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Ethical and Inclusive Use of Data Relating to Disabled People.md", "AGENTS.md", "README.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us", "title": "Nothing About Us Without Us", "text": "# Nothing About Us Without Us\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 1, "source_end_line": 2, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "d9db348180c165a8d1355e38aa90cf2eb472fa58725624a19c3c50fe77ddd158", "heading_path": ["Nothing About Us Without Us"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-inclusive-research-lived-experience-and-the-ethical-value-of-better-data", "title": "Inclusive research, lived experience and the ethical value of better data", "text": "## Inclusive research, lived experience and the ethical value of better data\n\n**Double FW Ltd · Version 1.0 · 6 November 2026** \n**Companion guidance for the Data Ethics and Disability repository**\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 3, "source_end_line": 7, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "5b2d6e2799360af7a12857f20531ad820653e92587c052ca094b5445aa39c6b7", "heading_path": ["Nothing About Us Without Us", "Inclusive research, lived experience and the ethical value of better data"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-1-start-with-accumulated-knowledge-then-involve-the-people-affected", "title": "1. Start with accumulated knowledge, then involve the people affected", "text": "## 1. Start with accumulated knowledge, then involve the people affected\n\nGovernance documentation, including Double FW guidance, WCAG, the European Accessibility Act (EAA) and other established frameworks, draws on accumulated knowledge from projects, research programmes, data sources, professional practice and expert contributions. These resources carry forward learning about disabled people's experiences. In that sense, knowledge “about us”, including knowledge contributed by people with lived experience, is already embedded in the foundation.\n\nThese sources have different purposes: WCAG is a technical standard, the EAA is legislation, and Double FW resources provide practical and organisational guidance. Their development processes and the extent of direct community participation differ. Recognising their inherited knowledge must not be taken as proof that every affected community has shaped every provision, or that following a framework fulfils “Nothing About Us Without Us”.\n\nThe foundation helps teams apply established knowledge without repeatedly asking participants to rediscover known barriers. It cannot tell a team everything about the people using a particular service, their priorities, the circumstances of use or how an experience makes them feel. WCAG itself recognises that it does not address every user need. Standards and lived experience should inform one another. [4, 5]\n\nMeaningful inclusion therefore requires current, inclusive research and a demonstrable opportunity for Deaf, disabled and neurodivergent people to influence project decisions. Their functional needs, emotional experiences, preferences, aspirations and disagreements belong within the evidence used to shape the whole service.\n\n**Applying established guidance is the starting point. Including people in research, interpretation and decisions makes that foundation relevant to the project.**\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 8, "source_end_line": 19, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "5b6a23df20f4268ac6b8aece3fec4f0da7b48f0ac9c0572c3e6ff2ee59e5c599", "heading_path": ["Nothing About Us Without Us", "1. Start with accumulated knowledge, then involve the people affected"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-2-how-this-guidance-fits-the-repository", "title": "2. How this guidance fits the repository", "text": "## 2. How this guidance fits the repository\n\nThis document extends the repository's existing principle of “Nothing About Us Without Us” into mainstream audience research, customer research and project governance. It uses the existing ethical framework for data collection and protection rather than creating a parallel set of rules.\n\n| Existing resource | Use it for | What this document adds |\n| --- | --- | --- |\n| [Ethical and Inclusive Use of Data Relating to Disabled People](./Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md) [1] | Equivalence, dignity, minimum necessary collection, aggregation, accessibility signals, research participation, AI use and governance | How inclusive recruitment and lived experience shape ordinary research and decisions within those safeguards |\n| [CUSTOMER RESEARCH FRAMEWORK](./CUSTOMER%20RESEARCH%20FRAMEWORK.md) [2] | Customer identification, jobs to be done, journeys, pain points, insight records and organisational intelligence | How to populate and challenge those structures using inclusive evidence, while distinguishing observed findings from assumptions |\n| [POUR for Branding Systems and CX](https://github.com/Double-FW/POUR-for-Branding-Systems-and-CX) [3] | Emotional outcomes and the relationship between accessibility, brand promise and experience | How to investigate feeling seen, capable, confident, safe and valued alongside functional outcomes |\n\nApply this guidance across research-led projects, including marketing, communications, product discovery, service design, content, pricing, support, retention and evaluation. Inclusion is relevant even when the research question does not mention accessibility.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 20, "source_end_line": 31, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "0c2378abe51a603c745a3ffaddd062e23b7b0eefcbefdc8181046219499acb87", "heading_path": ["Nothing About Us Without Us", "2. How this guidance fits the repository"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-3-include-people-in-ordinary-research", "title": "3. Include people in ordinary research", "text": "## 3. Include people in ordinary research\n\nDeaf, disabled and neurodivergent people are customers, colleagues, decision-makers and members of audiences. They have opinions about value, trust, language, aesthetics, convenience, quality and service priorities, as well as accessibility.\n\nRecruiting them only for specialist accessibility studies positions them outside the assumed mainstream audience. Systematically leaving them out of ordinary research is a form of **othering**: treating their perspectives as a separate concern rather than part of the population the service exists to serve. The omission can distort findings even when researchers have no intention to exclude.\n\nEvery project using qualitative audience or customer research should include Deaf, disabled and neurodivergent participants within its recruitment plan. This applies to interviews, concept research, diary studies, co-design, message testing, brand research and usability sessions. Participants should be able to discuss the same substantive questions as everyone else, with accessible methods and appropriate flexibility.\n\nDo not reinterpret everything a participant says as an accessibility issue. A Deaf participant's objection to pricing may be about pricing. A neurodivergent participant's preference for a brand may reflect its values. Ask what matters to the person and retain the context of their answer.\n\nSpecialist accessibility research remains valuable for exploring particular barriers or interactions in depth. It complements inclusive mainstream research.\n\n### Proportional representation and deliberate breadth\n\nAll audience and customer research programmes should plan proportional inclusion of people who self-identify as Deaf, disabled or neurodivergent, using a relevant audience benchmark. Apply this differently according to the method:\n\n| Research situation | Practical approach | Interpretation boundary |\n| --- | --- | --- |\n| Quantitative audience or customer research | Set recruitment targets using credible, dated evidence about the intended audience; check accessible participation and non-response; obtain sufficient responses for planned comparisons | A quota alone does not establish representativeness. Explain sampling, weighting, uncertainty and coverage limitations |\n| Small qualitative studies | Deliberately recruit a range of relevant lived experiences, functional needs and circumstances across the project's research rounds | Small samples cannot reproduce population proportions reliably or establish prevalence. Include people purposefully and document gaps |\n| Research into a particular barrier or underserved group | Recruit additional participants with relevant experience, alongside evidence from other affected groups | Extra recruitment is appropriate for learning about an issue; it does not establish that group's population share |\n| Services whose current customers reflect past exclusion | Consider intended customers, eligible non-users, people who abandoned the journey and people using alternatives | Current customer data may understate unmet demand and access barriers |\n\nUse voluntary self-identification, allow overlapping identities and provide “prefer not to say”. Deaf identity, disability and neurodivergence are not mutually exclusive, and people use these terms differently. A person may have significant functional needs without identifying with any of them. Do not require diagnosis or medical proof.\n\nRecord the benchmark, definitions, recruitment rationale and achieved coverage. Where relevant evidence is uncertain, state the uncertainty and improve the plan. Do not invent a universal percentage or treat one participant as representing an entire community. Consider intersections such as age, language, ethnicity, income, digital confidence and multiple impairments where they affect the research question.\n\nIf recruitment falls short, change channels, provide access support, extend recruitment or add a research round. Record unresolved gaps and the decisions they limit; do not quietly remove inclusion from the plan.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 32, "source_end_line": 60, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "9907c356a6a2dae581ec7018e15f2555398affb0b9fcae2f38f4fbb01421f382", "heading_path": ["Nothing About Us Without Us", "3. Include people in ordinary research"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-4-participation-must-have-influence", "title": "4. Participation must have influence", "text": "## 4. Participation must have influence\n\nRecruitment is necessary, but attendance alone does not establish meaningful participation. People need a realistic opportunity to shape what is investigated and what happens next.\n\n1. **Set the questions together.** Involve affected people or appropriately constituted community advisers early enough to challenge assumptions, identify missing outcomes and influence priorities.\n2. **Make participation accessible.** Offer accessible recruitment, consent and materials; suitable communication methods; interpreters or captions where needed; flexible timings, breaks and alternatives to live sessions. Ask about access arrangements without requiring a condition to be disclosed.\n3. **Respect people's contribution.** Budget fair payment, access support and expenses. Explain the role, effort, influence and limits before participation. Do not depend on unpaid advocacy to make the project inclusive.\n4. **Interpret findings with people.** Offer opportunities to challenge researchers' interpretations, especially where findings concern identity, dignity or emotional experience. Preserve different perspectives rather than manufacturing consensus.\n5. **Show the effect.** Record what changed, what remains unresolved, who made the decision and why. Return an accessible account of the findings and resulting actions to participants who want it.\n\nParticipants contribute expertise about their own experiences. They are not responsible for certifying a service, representing everyone with a similar identity, or resolving technical defects already covered by established guidance. Community advisers and representative organisations add valuable perspectives, while customers and non-customers contribute different evidence about actual and potential use. Follow the existing framework's **Customers Versus Research Participants** discussion when interpreting those differences. [1]\n\nUse established patterns and technical checks before evaluative sessions wherever possible. Early discovery and co-design should still involve people before a finished product exists; removing known defects from a prototype is different from postponing participation until the end.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 61, "source_end_line": 74, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "5874171549f3ff417df87b25f56fccf2178cc89b6bc14767d85d7a04be6b8c27", "heading_path": ["Nothing About Us Without Us", "4. Participation must have influence"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-5-lived-experience-can-reveal-wider-design-opportunities", "title": "5. Lived experience can reveal wider design opportunities", "text": "## 5. Lived experience can reveal wider design opportunities\n\nPeople with permanent conditions or impairments may have developed strategies, preferences and detailed knowledge that help teams recognise barriers in other circumstances. That expertise can expose design possibilities that a conventional sample misses.\n\nThis creates practical value as well as ethical value: richer evidence can improve decisions, reduce avoidable rework and reveal underserved demand. Inclusion can turn a commitment perceived as merely worthy into work that is demonstrably worthwhile. The rights and dignity of participants remain sufficient reasons to include them, whether or not a wider commercial benefit follows.\n\nThe examples below are **proposed research approaches**, not claims about completed studies or guaranteed outcomes.\n\n| Design context | Lived experience to include | Opportunities to investigate | What still needs validation |\n| --- | --- | --- | --- |\n| Audio-only interaction, such as an Alexa skill | Congenitally blind people, alongside other blind users and people with varied audio-interface experience | Discoverability without visual cues, spoken structure, memory demands, orientation, confirmation and recovery | Sighted people using audio-only services may have different skills, expectations and environmental constraints |\n| Important or life-changing decisions | Neurodivergent people with varied experiences, people with cognitive access needs and people who have faced the relevant decision | Clear options, manageable information, time to reflect, saved progress, reversible steps and support without loss of agency | Neurodivergence is not a proxy for distress or reduced capability; investigate the actual decision context |\n| Travel and disrupted journeys | Disabled travellers, neurodivergent travellers and people managing fatigue or unfamiliar environments | Predictability, disruption information, accessible alternatives, low-effort recovery and continuity across channels | Noise, movement, connectivity and urgency require evidence from realistic travel contexts |\n| Tax returns and forms with legal consequences | People with varied reading, memory, attention and executive-function needs, plus users with relevant experience | Plain-language explanations, consequences, error prevention, review, correction and confirmation | Domain knowledge and legal understanding differ from interaction skills; comprehension needs separate examination |\n| Judicial and medical situations | People with relevant lived experience, including Deaf, disabled and neurodivergent participants | Communication choice, dignity, privacy, informed understanding, pacing, support and trust | Power imbalance, distress and sensitive disclosure require appropriate methods and support |\n| Early morning, late evening or periods of low energy | People with fluctuating conditions, fatigue or relevant cognitive access needs | Resumability, reduced memory burden, flexible pace and recovery after interruption | Energy and cognition vary individually; time of day alone does not establish impairment |\n\nTreat transfer as a hypothesis: identify the original barrier, the strategy that helps, the new context and the evidence needed to test it. A successful design for one group may introduce difficulty for another. Validate major changes across affected populations, as required by the existing framework's **Accessibility Evidence Should Never Come From A Single Group** section. [1]\n\nDo not induce exhaustion, distress or unsafe distraction to reproduce a context. Use appropriately supported interviews, diaries, safe scenarios or observation chosen with participants. A disability simulation cannot replace lived experience.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 75, "source_end_line": 95, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "fe5356c5316d969851c7a64d418c4b4b8ab516f26bd6083916a5285bffba3c86", "heading_path": ["Nothing About Us Without Us", "5. Lived experience can reveal wider design opportunities"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-6-understand-capability-capacity-and-context-together", "title": "6. Understand capability, capacity and context together", "text": "## 6. Understand capability, capacity and context together\n\nFor this guidance, **capability** means what a person can do in the relevant interaction; **capacity** means the energy, attention, time and resources available to do it at that moment. These are working research terms, not clinical or legal assessments of capacity.\n\nA person who can understand a complex service in a quiet setting may find it difficult while travelling, experiencing pain, managing medication effects, under emotional pressure or making a consequential decision. The service can create disabling demands through ambiguity, unnecessary complexity, urgency or lack of recovery options.\n\nExplore the interaction between the person, the task and the circumstances. Record only context necessary for the research question. Do not assume that everyone is impaired early in the morning or late at night, that neurodivergent people have low capability, or that a diagnosis predicts preference or satisfaction.\n\nAsk participants what they were trying to do, what resources they had available, what the service demanded and what would have helped. Keep the explanation grounded in evidence about that interaction.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 96, "source_end_line": 105, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "39d1aebbf7f2ac7c7bb1166d1b7253444e680233491081f4c8ccc30eba69d6a7", "heading_path": ["Nothing About Us Without Us", "6. Understand capability, capacity and context together"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-7-use-functional-segmentation-without-building-disability-profiles", "title": "7. Use functional segmentation without building disability profiles", "text": "## 7. Use functional segmentation without building disability profiles\n\nThe existing ethical framework recommends alignment with functional accessibility categories reflected in EAA Annex I. **Section VII's functional performance criteria** provide a useful starting point for checking research coverage. They are not a prescribed research segmentation scheme, a medical classification or a complete account of lived experience. [1, 6]\n\nThe following are concise research prompts based on its eleven areas; they do not reproduce the legal requirements:\n\n| Section VII area, paraphrased | Research coverage prompt |\n| --- | --- |\n| Vision absent or limited; colour perception unavailable | Can people obtain and use information through the sensory modes available to them? |\n| Hearing absent or limited | Can people follow communication through appropriate alternatives and audio options? |\n| Vocal input unavailable | Can people participate without speech or other vocal input? |\n| Manipulation, strength or reach limited | Can people interact without unsuitable physical demands? |\n| Photosensitive seizure risk | Does evaluation avoid hazardous visual stimulation? |\n| Cognitive demands | Can people understand, decide, remember and recover with suitable support? |\n| Privacy when using accessibility features | Can people use access options without unnecessary disclosure? |\n\nFor detailed functional coverage, distinguish each applicable area within these grouped prompts. Add overlapping needs, communication language, task familiarity, emotional experience and context where relevant. People may belong to several functional segments and their needs may change.\n\nThis use of Section VII is a research design adaptation. The EAA gives those criteria a particular legal role; it does not prescribe the recruitment targets, satisfaction measures or segmentation proposed here. Inclusive research findings also do not establish legal conformity.\n\nKeep three dimensions distinct: **self-described identity**, **functional need or preference**, and **context of use**. Collect each only when it serves a defined purpose. Do not infer identity or diagnosis from assistive technology, captions, navigation behaviour or accessibility settings. Refer to **Understanding Accessibility Signals** and **Disability Must Never Become a Justification for Surveillance** in the existing framework. [1]\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 106, "source_end_line": 127, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "c034d658d5c4f4e75ff8d9b445ef7aaba3440750afab52b06781f5b7be82363a", "heading_path": ["Nothing About Us Without Us", "7. Use functional segmentation without building disability profiles"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-8-measure-satisfaction-in-the-context-of-the-service", "title": "8. Measure satisfaction in the context of the service", "text": "## 8. Measure satisfaction in the context of the service\n\nTask completion alone cannot establish equivalence. Investigate whether people receive the information they need, retain agency, expend reasonable effort and experience comparable dignity, confidence and satisfaction.\n\nUse Double FW's emotional reframing of POUR to frame research questions. The first four feelings correspond to its interpretation of POUR; **feeling valued** is an additional proposed emotional principle, not a fifth WCAG principle. [3]\n\n| Emotional outcome | Example question | Supporting evidence |\n| --- | --- | --- |\n| Feeling seen — Perceivable | “Was the information available in a form that worked for you?” | Information access, representation and missed content |\n| Feeling capable — Operable | “Could you do what you intended, with the level of support you chose?” | Completion, effort, independence and recovery |\n| Feeling confident — Understandable | “Did you understand your options, the consequences and what would happen next?” | Comprehension, uncertainty and confidence in decisions |\n| Feeling safe — Robust | “Could you rely on the service and your chosen tools throughout?” | Reliability, continuity, privacy concerns and reported trust |\n| Feeling valued — additional emotional principle | “Did you feel heard, respected and intentionally considered?” | Satisfaction, dignity, belonging and responsiveness |\n\nAdapt questions with participants; these are prompts, not a validated psychometric scale. Ask open questions before introducing labels that might lead an answer. Investigate feeling heard through both the service experience and the response to research contributions.\n\nSatisfaction must reflect the service's purpose. A tax return, court process or medical appointment should support clarity, dignity, agency and trust; an upsetting underlying outcome does not automatically mean the interface failed. Distinguish satisfaction with the process from satisfaction with the substantive outcome.\n\nExamine outcomes by relevant functional need and context where sample size and privacy permit. Report missing responses and abandonment alongside completed journeys. An overall average can conceal exclusion, while small subgroup differences may be too uncertain to interpret. Combine quantitative results with qualitative explanations and record unresolved barriers. Positive satisfaction scores cannot justify avoidable harm, loss of rights or failure to meet applicable requirements.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 128, "source_end_line": 147, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "03ffffa22404b4dde8c5380f4ad8a11907fff107f16c1c604327e5533ab84926", "heading_path": ["Nothing About Us Without Us", "8. Measure satisfaction in the context of the service"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-9-gather-richer-insight-with-minimum-necessary-data", "title": "9. Gather richer insight with minimum necessary data", "text": "## 9. Gather richer insight with minimum necessary data\n\nInclusive research does not require a permanent disability register. Apply the existing framework's collection, aggregation, consent, sharing and AI governance principles throughout. [1]\n\n- Separate recruitment and access-arrangement details from research findings wherever practical. Restrict access to each according to purpose.\n- Ask about functional needs and preferences directly when appropriate. Make identity disclosure optional and explain why any such question is asked.\n- Explain participation, recording, withdrawal arrangements and their limits, retention, access and intended uses in accessible terms. Participation consent does not automatically authorise publication, AI training or unrelated reuse.\n- Remove unnecessary identifiers early. Pseudonymisation does not make data anonymous; small groups, quotations and combinations of attributes can still identify someone.\n- Publish or share insight at a safe level of aggregation. Combine or suppress risky breakdowns without implying that the affected experience did not exist.\n- Do not use recordings, transcripts or sensitive responses in external AI tools or shared datasets without the permissions and safeguards required by the existing framework.\n\nWhere the CUSTOMER RESEARCH FRAMEWORK proposes personas, knowledge graphs, RAG records or training data, treat these as possible outputs subject to the ethical framework. A suggested output format is not permission to reuse participant data. Mark inferred personas, synthetic examples and hypotheses clearly; they are not observed testimony or substitutes for participants. [1, 2]\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 148, "source_end_line": 160, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "3b318af42da54184eba30603a787b4421b5c3ba7f4a95f4cdeca8042a4114d11", "heading_path": ["Nothing About Us Without Us", "9. Gather richer insight with minimum necessary data"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-10-turn-inclusive-evidence-into-evolving-guidance", "title": "10. Turn inclusive evidence into evolving guidance", "text": "## 10. Turn inclusive evidence into evolving guidance\n\nGuidance should evolve when better evidence reveals missing needs, failed assumptions or more effective approaches. Teams need a traceable path from research to change:\n\n1. Record the observation and relevant context, including contradictory experiences and evidence limitations.\n2. Review its interpretation with affected participants or advisers where appropriate.\n3. Identify whether it concerns local implementation, a reusable pattern or a gap in guidance.\n4. Assign a decision owner and agree an action, outcome measure and review point.\n5. Validate the proposed change across relevant groups and contexts, retaining established technical checks.\n6. Update local guidance, patterns or governance with a dated rationale and a privacy-safe evidence reference.\n7. Explain the result to contributors and monitor the experience after implementation.\n\nResearch may justify strengthening guidance or proposing changes to external standards. It does not allow a local project to rewrite WCAG or waive applicable legal duties. Where evidence conflicts, record the trade-off and seek an inclusive solution rather than allowing the largest group's preference to erase a smaller group's needs.\n\nUse the CUSTOMER RESEARCH FRAMEWORK's insight catalogue to connect findings to action. Add the following fields when useful, without storing identifiable participant detail in broadly accessible records. [2]\n\n| Field | Purpose |\n| --- | --- |\n| Research question and service context | Explain the decision the evidence can inform |\n| Recruitment coverage and gaps | Show whose perspectives were included and missing |\n| Functional and emotional outcomes | Retain both ability to participate and quality of experience |\n| Evidence reference and confidence | Distinguish observation, interpretation and hypothesis |\n| Participant or adviser interpretation | Record agreement, disagreement and alternative explanations |\n| Decision, owner and rationale | Make influence and accountability visible |\n| Validation and guidance revision | Show whether the change worked and what learning was retained |\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 161, "source_end_line": 186, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "7181a310ed50d074ec9f30dc4b0c4f81bd172ca3a326a8e2a646813caef61cb7", "heading_path": ["Nothing About Us Without Us", "10. Turn inclusive evidence into evolving guidance"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-11-project-review-checklist", "title": "11. Project review checklist", "text": "## 11. Project review checklist\n\nBefore treating research as sufficient evidence for a decision, check:\n\n- [ ] Established guidance has informed the work, and its limits are understood.\n- [ ] Mainstream research recruitment includes Deaf, disabled and neurodivergent people, even where accessibility is not the topic.\n- [ ] Quantitative targets have a relevant benchmark; qualitative recruitment has a deliberate breadth rationale.\n- [ ] Intended users and excluded non-users have been considered alongside current customers.\n- [ ] Participation is accessible, voluntary, appropriately supported and fairly paid.\n- [ ] People can influence questions, interpretations and decisions, with a clear account of what changed.\n- [ ] Functional needs, identity and context are distinguished without unnecessary disclosure or profiling.\n- [ ] Transferable insights have been tested in the proposed new context and across affected groups.\n- [ ] Evidence covers informational and emotional equivalence, satisfaction and dignity as well as completion.\n- [ ] Data collection, reuse, sharing and AI processing follow the existing ethical framework.\n- [ ] Coverage gaps, uncertainty, disagreement and remaining barriers are explicit.\n- [ ] Findings have an owner, an action and a route into improved guidance and future evaluation.\n\nIf evidence is insufficient, record what the project can reasonably conclude, what it cannot conclude and the further involvement required. Inclusion is demonstrated through participation that changes understanding and action, supported by ethical data practices and continuing accountability.\n\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 187, "source_end_line": 205, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "8dc7d8292dc569345082de673dff9cffab7c6c035394e1f09659ee689a14a29c", "heading_path": ["Nothing About Us Without Us", "11. Project review checklist"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "Nothing_About_Us_Without_Us.md::nothing-about-us-without-us-references-and-document-relationships", "title": "References and document relationships", "text": "## References and document relationships\n\nRepository sources were reviewed on 6 November 2026. Sections 3–11 develop practical recommendations from the user's brief and the repository's principles; they should not be read as quotations from, or additional legal requirements imposed by, external standards.\n\n1. Double FW, [Ethical and Inclusive Use of Data Relating to Disabled People](./Ethical%20and%20Inclusive%20Use%20of%20Data%20Relating%20to%20Disabled%20People.md). Primary ethical framework; especially Nothing About Us Without Us, Customers Versus Research Participants, Inclusive Data Is Broader Than Disability, Data and the European Accessibility Act, AI Training and Accessibility Data, and Governance Principles.\n2. Double FW, [CUSTOMER RESEARCH FRAMEWORK](./CUSTOMER%20RESEARCH%20FRAMEWORK.md). Companion structure for customer intelligence and insight records.\n3. Double FW, [The WCAG POUR principles reframed through emotional outcomes for branding systems](https://github.com/Double-FW/POUR-for-Branding-Systems-and-CX/blob/main/The%20WCAG%20POUR%20principles%20reframed%20through%20emotional%20outcomes%20for%20branding%20systems.md). Source of the emotional mapping used here.\n4. W3C, [WCAG 2.2](https://www.w3.org/TR/WCAG22/), Abstract and Introduction. Scope and limitations of the shared technical foundation.\n5. W3C WAI, [Involving Users in Web Projects for Better, Easier Accessibility](https://www.w3.org/WAI/planning/involving-users/). Supports combining standards with early and continuing involvement and recognising diversity within disability groups.\n6. European Union, [Directive (EU) 2019/882](https://eur-lex.europa.eu/eli/dir/2019/882/oj), Annex I, Section VII. Source for the functional coverage prompts; the research adaptation is explained in section 7.\n\n**Repository placement:** Save this file in the repository root so its relative links resolve. Suggested README entry: `Nothing_About_Us_Without_Us.md — companion guidance on inclusive mainstream research, lived experience, functional and emotional outcomes, and participation that influences decisions.`\n", "record_type": "participation_and_research_governance_guidance", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "Nothing_About_Us_Without_Us.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/Nothing_About_Us_Without_Us.md", "source_start_line": 206, "source_end_line": 217, "source_document_sha256": "dffe920183b976178f2deb04f2b566dc26b995224b454708158076d139a0cd25", "text_sha256": "78572f4bb090f63d150dd3d0886e3d878caf8ee34751d8a4ef95026a827f233c", "heading_path": ["Nothing About Us Without Us", "References and document relationships"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Nothing About Us Without Us\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["Nothing_About_Us_Without_Us.md", "AGENTS.md", "README.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "README.md::data-ethics-and-disability", "title": "Data Ethics and Disability", "text": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "record_type": "orientation", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "README.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/README.md", "source_start_line": 1, "source_end_line": 4, "source_document_sha256": "50a3fa33d1158d2d53a43bbdd884fd6f9e07b54654cf00d727ae6663e1bc2fc7", "text_sha256": "3c2fd42e72b3b2a04e7f82732f2a731a39909f053c6422fcddb0a6508045a9c0", "heading_path": ["Data Ethics and Disability"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["README.md", "AGENTS.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "README.md::data-ethics-and-disability-what-this-repository-covers", "title": "What this repository covers", "text": "## What this repository covers\n\nThis repository brings together practical guidance for using data related to disabled people in a fair, privacy-preserving, and dignity-respecting way. It is designed to support organisations that need to understand accessibility barriers and improve outcomes without turning people into profiles or treating accessibility as a reason for surveillance.\n\nIt includes:\n\n- `Ethical and Inclusive Use of Data Relating to Disabled People.md` — core principles for dignity, privacy, minimum necessary data, equivalence, and governance.\n- `Nothing_About_Us_Without_Us.md` — how to include disabled and neurodivergent people in ordinary research and decision-making.\n- `Assistive_Technology_Data_Flows_Risks_and_Review.md` — review patterns for assistive technology, AI tools, third-party scripts, forms, and other data flows.\n- `CUSTOMER RESEARCH FRAMEWORK.md` — a structured framework for research, personas, journeys, pain points, and insight capture.\n\n", "record_type": "orientation", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "README.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/README.md", "source_start_line": 5, "source_end_line": 15, "source_document_sha256": "50a3fa33d1158d2d53a43bbdd884fd6f9e07b54654cf00d727ae6663e1bc2fc7", "text_sha256": "e62767f5f50a02e30cab082a88e69a1bffc58bcf0e1a412c7160af8b772cd04e", "heading_path": ["Data Ethics and Disability", "What this repository covers"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["README.md", "AGENTS.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "README.md::data-ethics-and-disability-processes-this-repo-can-support", "title": "Processes this repo can support", "text": "## Processes this repo can support\n\nThis guidance can inform and strengthen:\n\n- accessibility and inclusion reviews\n- customer and user research design\n- AI procurement and governance checks\n- HR and employee-facing process review\n- product and service design decisions\n- data governance and privacy reviews\n- vendor and tool assessment for assistive technology or accessibility features\n- policy and operating model design for equitable digital experiences\n\n", "record_type": "orientation", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "README.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/README.md", "source_start_line": 16, "source_end_line": 28, "source_document_sha256": "50a3fa33d1158d2d53a43bbdd884fd6f9e07b54654cf00d727ae6663e1bc2fc7", "text_sha256": "a259cd593bf80c2c18f09cd7b29245360d1147bc4fbcc978f5c9a42badc100f4", "heading_path": ["Data Ethics and Disability", "Processes this repo can support"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["README.md", "AGENTS.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "README.md::data-ethics-and-disability-core-principles", "title": "Core principles", "text": "## Core principles\n\n- Nothing about us without us.\n- Equivalence of experience over narrow task completion.\n- Dignity as an accessibility requirement.\n- Minimum necessary data and privacy by design.\n- Aggregate insight instead of individual profiling.\n- Human accountability and evidence-based decisions.\n\n", "record_type": "orientation", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "README.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/README.md", "source_start_line": 29, "source_end_line": 37, "source_document_sha256": "50a3fa33d1158d2d53a43bbdd884fd6f9e07b54654cf00d727ae6663e1bc2fc7", "text_sha256": "d1e284e584a983afe88ea0002217c60c233a0c2c58530a3f5d039fc6ba522614", "heading_path": ["Data Ethics and Disability", "Core principles"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["README.md", "AGENTS.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "README.md::data-ethics-and-disability-related-double-fw-repositories", "title": "Related Double-FW repositories", "text": "## Related Double-FW repositories\n\n- [Inclusion and Accessibility Policy](https://github.com/Double-FW/Inclusion-and-Accessibility-Policy) — policy development, governance, and accountability.\n- [AI Accessibility Maturity Model](https://github.com/Double-FW/AI-Accessibility-Maturity-Model) — organisational maturity across data, AI, ethics, and inclusion.\n- [The Little Book of Accessibility](https://github.com/Double-FW/The-Little-Book-of-Accessibility) — broader accessibility guidance.\n- [Executive and Organisational Commitments](https://github.com/Double-FW/Executive-and-Organisational-Commitments) — disability inclusion commitments and organisational accountability.\n- [Accessibility Champions](https://github.com/Double-FW/Accessibility-Champions) — establishing and governing an accessibility champions network.\n- [POUR for Branding Systems and CX](https://github.com/Double-FW/POUR-for-Branding-Systems-and-CX) — emotional outcomes and accessibility in branding and customer experience.\n\n", "record_type": "orientation", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "README.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/README.md", "source_start_line": 38, "source_end_line": 46, "source_document_sha256": "50a3fa33d1158d2d53a43bbdd884fd6f9e07b54654cf00d727ae6663e1bc2fc7", "text_sha256": "bf11081c009ccc3f6789d10088d4928c9e9caf7d3718960efbbf4547f6d4e2d3", "heading_path": ["Data Ethics and Disability", "Related Double-FW repositories"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["README.md", "AGENTS.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |
| {"id": "README.md::data-ethics-and-disability-license", "title": "License", "text": "## License\n\nThis project is licensed under the MIT License. See `LICENSE` for details.\n", "record_type": "orientation", "source_repository": "Double-FW/Data_Ethics_and_Disability", "source_path": "README.md", "source_commit": "77163fa48b6df5839e1294ec02797abf67c73980", "source_url": "https://github.com/Double-FW/Data_Ethics_and_Disability/blob/77163fa48b6df5839e1294ec02797abf67c73980/README.md", "source_start_line": 47, "source_end_line": 49, "source_document_sha256": "50a3fa33d1158d2d53a43bbdd884fd6f9e07b54654cf00d727ae6663e1bc2fc7", "text_sha256": "af7bcb0a475a1caa7a1349b8f4790bc3ba1cdcf5ae4281956a5ae63968ef0255", "heading_path": ["Data Ethics and Disability", "License"], "metadata_json": "{}", "declared_source_license": "not_declared_in_document_front_matter", "repository_license": "MIT", "source_status": "see_source", "source_document_intro": "# Data Ethics and Disability\n\nThe ethical use of disabled customer data.\n\n", "dataset_version": "1.0.0", "source_commit_date": "2026-11-10", "language": "en-GB", "required_context_paths": ["README.md", "AGENTS.md", "Ethical and Inclusive Use of Data Relating to Disabled People.md", "Nothing_About_Us_Without_Us.md", "Assistive_Technology_Data_Flows_Risks_and_Review.md"], "review_status": "faithful_extraction_checked_not_privacy_legal_or_research_approval"} | |