A starter checklist for AI literacy evidence.
This iFeed checklist helps teams convert Article 4 into practical evidence-readiness questions. It is designed for regulated life sciences and health-tech settings where AI use may touch clinical, quality, regulatory, digital-health, or operational decisions.
EU AI Act Article 4 AI Literacy Starter Assessment
A fillable iFeed evidence-readiness checklist for regulated life sciences and health-tech teams. Complete the details below before using the checklist so the exported PDF carries your context, ownership, and review record.
Checklist table
Use the status column to mark readiness, then record evidence notes and ownership/actions. The table header repeats across PDF pages and the completed fields are retained in the downloaded PDF.
| Checkpoint | Question | Suggested evidence | Status | Evidence notes | Owner / action |
|---|---|---|---|---|---|
| AI system inventory | List AI systems or AI-enabled workflows used, provided, deployed, evaluated, or relied on by the team. | Inventory owner, system name, use context, users, affected persons. | |||
| Role mapping | Identify who deals with operation or use of each AI system on behalf of the organisation. | Role list, responsibilities, access level, decision involvement. | |||
| Literacy baseline | Assess current skills, knowledge, training, and practical understanding for each role. | Training record, competency evidence, interview notes, role matrix. | |||
| Use-context fit | Check whether literacy is appropriate for the actual use context, not only the tool category. | Workflow map, SOP link, risk notes, task description. | |||
| Risk and affected persons | Document how AI output could affect patients, trial participants, users, staff, or regulated decisions. | Impact notes, risk register link, affected-person map. | |||
| Human oversight | Define what users must understand before accepting, challenging, escalating, or overriding AI output. | Oversight instruction, escalation route, reviewer role. | |||
| Data and limitation awareness | Ensure users understand data limitations, model limitations, output uncertainty, and inappropriate use. | Training content, limitation summary, known-use restrictions. | |||
| Vendor and tool evidence | Capture what the vendor or internal developer provides about intended use, limitations, and training needs. | Vendor documentation, release note, user guide, validation summary. | |||
| GxP / regulated workflow link | Mark whether the AI system touches GxP, clinical, quality, regulatory, safety, or device workflows. | Workflow classification, QA review note, business owner. | |||
| Records and review cycle | Define how literacy evidence will be kept current as systems, users, or obligations change. | Review date, owner, version history, update trigger. | |||
| Gap and action log | Record gaps between current literacy and what the use context reasonably requires. | Gap description, priority, action owner, target date. | |||
| Management visibility | Create a route for leadership, QA, regulatory, or governance review where risk justifies it. | Review minutes, sign-off note, decision record. |
Source trace
Regulation (EU) 2024/1689.
Official Journal text for the Artificial Intelligence Act.
European Commission AI literacy questions and answers.
Official Q&A used to keep public wording cautious and bounded.
Article 4 source object.
Stable iFeed source page connecting Article 4 to interpretation and evidence-readiness.