Glossary
Updated
What does AI literacy mean?
AI literacy is the working knowledge someone needs to use, buy or oversee AI responsibly. It covers what a system is for, how it produces its outputs, where it tends to go wrong and what happens to the data people give it. The EU AI Act defines it in Article 3(56) as the skills, knowledge and understanding that let providers, deployers and affected persons make an informed deployment of AI systems, and gain awareness of AI's opportunities and risks and the harm it can cause.
The right level depends on the role. A claims handler reading an AI summary needs to know when to check the source documents. A developer building on a model needs to understand its known limitations. A board member needs enough to ask the right questions about risk.
What does the EU AI Act require?
Article 4 has applied since February 2, 2025, to providers and deployers alike. As first written, it asked them to take measures to ensure a sufficient level of AI literacy among their staff and others running AI on their behalf. The Digital Omnibus, which entered into force on July 27, 2026, softened that duty. Article 4 now asks providers and deployers to take measures to support the AI literacy of their staff and anyone running AI for them, suited to their role. It no longer asks them to ensure a sufficient level, and they needn't guarantee any specific level for any individual. The Commission and member states must help, particularly smaller firms.
The duty is softer, but it still applies. Our EU AI Act guide sets it alongside the Act's other dates and duties.
How do other frameworks treat it?
ISO/IEC 42001 covers the same ground through its competence and awareness clauses, 7.2 and 7.3. They ask an organization to decide what competence each AI role needs, provide it and keep evidence. The NIST AI Risk Management Framework asks, under GOVERN 2.2, that personnel and partners receive AI risk management training so they can carry out their duties. Neither sets a curriculum, so the content is yours to decide.
How do you meet the AI literacy duty?
Start from the AI people use. Knowing which tools are in use across the company tells you which training matters first: a team using a writing assistant needs different guidance from a team building agents. Match training to each role, and keep a record of who completed what and when. Update it when the tools change, because a new model or feature can change what people need to know. An acceptable use policy gives the training a set of rules to point to.
Enzai doesn't run training. Its AI discovery and inventory shows which AI tools are in use and lets you give each one an owner, so you can aim training at the tools people rely on.
Example
A European insurer's claims team uses an approved AI tool to summarize medical reports. To meet Article 4, the insurer gives claims handlers a short module on where summaries drop detail and when to read the full report, and gives their managers a session on reviewing AI-assisted decisions. It records completion in its learning system and reviews the module whenever the tool's model changes. When a regulator asks, it can show which roles were trained, on what, and when.