Article 4 — AI literacy
What Article 4 requires
Article 4 of the EU AI Act obliges both providers and deployers of AI systems to ensure, to their best extent, a sufficient level of AI literacy among their staff and any other persons who operate or use AI systems on their behalf — including contractors, agency workers, and partners with access to the system.
The obligation is not confined to high-risk systems. It applies horizontally, across every AI system within the scope of the Regulation, regardless of risk classification.
In force since 2 February 2025
Article 4 has applied since 2 February 2025, alongside the prohibitions in Article 5. It stands independently of the staggered timeline for high-risk obligations under Chapter III.
Under the Regulation as currently in force, the Annex III high-risk obligations apply from 2 August 2026. The political agreement on the so-called "Digital Omnibus" reached on 7 May 2026 would postpone those Annex III obligations to 2 December 2027 — but that postponement is not yet formally adopted. Until the amending act enters into force, 2 August 2026 remains the operative date. Either way, Article 4 is unaffected: it already applies.
In practice: an organisation that deploys AI in HR processes today is already within scope of Article 4, even if its conformity assessment work for any high-risk systems it operates still lies ahead.
What "sufficient literacy" means in practice
The Regulation does not prescribe a single curriculum. Article 3(56) defines AI literacy as the skills, knowledge and understanding needed to make an informed deployment of AI systems and to be aware of the opportunities, risks and possible harms they entail.
In operational terms, supervisory authorities and the AI Office expect literacy measures to be:
- Role-appropriate — calibrated to what each person actually does with the system (configure, operate, review outputs, make decisions, audit).
- Risk-appropriate — deeper for staff handling use cases that qualify as high-risk under Annex III (for example, recruitment screening or employee evaluation, subject to the Article 6(3) derogation).
- Context-aware — reflecting the technical knowledge, experience, education and training already held by the recipient, and the population on which the system is used.
- Documented — recorded in a form that can be produced on request, with dates, content, and attendance.
Generic "AI awareness" videos do not, on their own, satisfy Article 4 for staff operating systems that affect employment decisions.
Why this matters for HR-tech
HR-tech sits at the intersection of two pressures.
First, many HR use cases fall within Annex III point 4 — AI systems intended to be used for the recruitment or selection of natural persons; for decisions affecting the terms of work-related relationships, promotion or termination; for task allocation based on individual behaviour or personal traits; and for monitoring and evaluating the performance and behaviour of employees. Whether a specific system is in fact high-risk also depends on Article 6(3): an Annex III system is not high-risk where it does not pose a significant risk of harm — for example, because it performs a narrow procedural task, improves the result of a previously completed human activity, or detects patterns without replacing or influencing the human assessment — provided it does not perform profiling, which always triggers high-risk status.
Second, HR-tech is typically embedded in customer workflows, which means literacy obligations cascade: the provider must ensure its own staff are literate, and the deployer (the customer) must ensure the same of theirs.
For HR-tech vendors, this creates a concrete commercial expectation. Enterprise procurement teams, works councils, and DPOs increasingly ask vendors to demonstrate not only that their system is compliant, but that the customer's operating staff will be enabled to use it competently. Article 4 evidence has become part of the sales conversation.
How PowerQuant delivers Article 4 evidence
PowerQuant produces two artefacts that map directly to Article 4 expectations:
- AI literacy register — a structured record of the roles that interact with the AI system, the literacy level required for each role, the measures in place, and the responsible owner. Designed to be maintained as staff and use cases change.
- Training documentation pack — module outlines, attendance records, assessment evidence, and refresh schedules, tied to the roles in the register. Includes a vendor-side pack for your own staff and a deployer-side pack that customers can adopt or adapt.
Both artefacts are delivered as part of the evidence package, versioned, and cross-referenced from the broader technical documentation, so that an auditor or customer compliance team can trace the chain from regulation, to role, to person, to training event.
If Article 4 is on your roadmap and you would like to see the register and training templates we ship, request a sample pack.