Artificial Intelligence
AI training: the Article 4 obligation of the AI Act for occupational health services and social care providers
Your teams already use AI. Which uses should training cover, which mistakes should staff be able to avoid, and how do you choose a useful programme? Start here.
Need to apply this to your organisation?
Explore AI literacy trainingAn occupational physician using AI to prepare notes, a medical secretary using an assistant and a care home exploring a chatbot face different risks. Article 4 of the AI Act requires AI providers and deployers to support the development of AI literacy for people acting on their behalf. As amended by Regulation (EU) 2026/1744, it expressly does not require any specific level of AI literacy to be guaranteed for each person. Read the consolidated text.
Occupational health services and social and medico-social care providers are doubly concerned: because their teams are adopting these tools quickly, and because they use them on the most protected data there is — health data covered by professional secrecy.
What Article 4 actually says
The amended Article 4 requires providers and deployers to take measures supporting AI literacy, considering users’ technical knowledge, experience, existing education and training, and the use context. It does not mandate a single universal course or individual certificate. The European Commission explains the scope.
Three points matter. First, the measures must be proportionate to the role, system and risks, so different teams may need different approaches. Second, training can be appropriate but is not an automatically mandated certification course. Third, Article 4 has applied since February 2025, while its wording was updated in July 2026 and other AI Act requirements have different effective dates.
What a serious “Article 4” training programme must cover
A generic one-hour awareness presentation may be insufficient for higher-risk use cases, but the AI Act sets no universal minimum course duration. A relevant occupational health or social-care programme could address:
- The useful basics: what an AI system can do, cannot do, and why it gets things wrong (bias, hallucinations, limits of training data);
- The legal framework applied to the job: the AI Act (risk categories, human oversight), the GDPR and professional secrecy — what may be entered into which tool, and what must never be;
- The team's real use cases, reviewed one by one: which are allowed, under what conditions, with what human verification;
- Governance reflexes: who to notify about a new tool, how to escalate a doubt or an incident, where to find the internal AI-use policy.
The organisation should document its selected measures, training programme, participants, learning materials and follow-up where appropriate. Attendance alone is neither a legal compliance certificate nor proof that every relevant AI system meets all applicable requirements.
Getting compliant: four steps
Step 1 — map actual uses, including “unofficial” uses of consumer tools: that is often where the risk sleeps. Step 2 — write a short, workable AI-use policy: authorised tools, prohibited data, human verification, reporting. Step 3 — train by profile: a common foundation for everyone, a reinforced module for healthcare professionals and roles handling sensitive data. Step 4 — document and update: a register of uses, proof of training, an annual review — tools change fast, and the training must keep up.
This is precisely the format of our “Mastering AI (Article 4)” training, designed for healthcare organizations and adapted to each structure — with your real use cases as the raw material.
Explore further: the rules, decisions and sources
Why occupational health services are on the front line
An occupational health service concentrates everything that makes the question sensitive: occupational health data covered by medical secrecy, significant volumes (thousands of monitored employees), multidisciplinary teams with very different profiles — occupational physicians, nurses, assistants, prevention specialists, support functions — and rapid adoption of AI-enabled tools: drafting assistance for reports, pre-analysis of questionnaires, prioritisation of medical visits, consultation transcription.
A tailored programme may help build those practical safeguards alongside organisational policies, technical controls and human oversight. Article 4 does not prescribe one universal training format; GDPR requirements still apply to the processing of personal data.
And social care providers: same tools, more vulnerable people
In social and medico-social care — care homes, home-care services, disability services, child protection — AI arrives through user records, schedules, handovers, and sometimes monitoring or assistance devices. The people concerned are often vulnerable, which raises the ethical and legal bar another notch.
The Article 4 literacy duty already applies. Other AI Act provisions have different commencement dates, some affected by the 2026 amendments. Teams should verify the consolidated timeline for the particular system and legal role rather than assume that every high-risk requirement is already in force.
The questions we are asked
“Does Article 4 apply without high-risk AI?” It is not limited to high-risk categories but concerns providers and deployers of AI systems. “Must everyone pass the same course?” No: there is no single mandatory programme or universal individual certificate. Measures should fit the context. “Is missing training subject to its own fine?” Article 4 does not provide a stand-alone course-related fine; the relevant supervision framework and other obligations must be assessed separately. “How often?” No universal interval is prescribed; organisations should review measures when tools, roles or risks change.
Official sources
Receive our next insightsPractical analysis of health data and AI.
Share this article
General information. Applying it to your situation requires a review of your project.