Guide · Last reviewed 26 August 2026

AI literacy under Article 4: what companies actually have to do

Article 4 is the EU AI Act's quietest obligation — in force since 2 February 2025, applying to virtually every company that provides or uses AI, and routinely misquoted. The Digital Omnibus (Regulation (EU) 2026/1744) softened its wording: companies must take measures to support the development of AI literacy of staff and other persons operating AI systems on their behalf. They do not have to guarantee that every person reaches a defined competence level.

What counts as a measure

The obligation is context‑dependent: measures should take into account staff members' technical knowledge, experience and training, and the context the AI systems are used in. In practice, companies evidence Article 4 with a short written policy, role‑appropriate training, and attendance records. None of these formats is mandated by the Act — they are simply the auditable way to show measures exist.

Who it covers

Both providers and deployers, from the day they provide or use any AI system — the duty is not limited to high‑risk AI. That makes Article 4, together with the Article 50 transparency duties, the compliance floor that applies to ordinary companies using everyday AI tools today.

How RegShelf covers it

The kit includes a signature‑ready AI literacy policy, a training deck, and an attendance log — with recommended cadences clearly labelled as internal controls, not statutory deadlines.