What does the EU AI Act require for AI literacy?

Every provider and every deployer of an AI system, whatever the risk class. A company whose staff use a general-purpose chatbot, a writing assistant or an AI feature in a business tool is a deployer of that system. The duty covers its staff and other persons dealing with the operation and use of AI systems on its behalf, such as contractors.

The original text asked providers and deployers to ensure, to their best extent, a sufficient level of AI literacy. Regulation (EU) 2026/1744, in force since 27 July 2026, replaced Article 4: providers and deployers now take measures to support the development of AI literacy, without having to guarantee any specific level for any individual, and the Commission and the member states support those efforts. The obligation itself stays, and national authorities supervise it.

• Know your AI use: keep a register of the AI systems your teams build, buy or use, and who uses them.
• Match the training to the role: general awareness for every user, deeper training for people who configure systems, oversee them or decide on their output.
• Set rules: an acceptable-use policy for AI tools, read and acknowledged by staff.
• Keep records: who was trained or acknowledged which version, and when.
• Review: update the measures when you add systems or the use changes.

For high-risk systems, Article 26(2) separately requires deployers to assign human oversight to people with the necessary competence, training and authority.

In AI Governance in Acuna GRC, the AI system register shows which systems you use and who they affect, and each overseer assignment records a competence basis. In the Staff Portal, you publish your AI use policy and staff read and acknowledge it, with campaigns and reminders, and each acknowledgement is kept with the version read. See AI Governance in Acuna GRC.