AI literacy is the ability to understand how AI systems work, use them responsibly, and judge their outputs and societal effects. Under EU AI Act Article 4 it is also a legal duty for providers and deployers of AI systems (see below).
Legal anchor: EU AI Act Article 4 (as amended by the Digital Omnibus)
- Article 4 of Regulation (EU) 2024/1689 has applied since 2025-02-02. The Digital Omnibus on AI (Regulation (EU) 2026/1744, in force 2026-07-27 per ai-regulation-and-policy) rewrote it from an outcome duty into an obligation of means: providers and deployers must “take measures to support the development of AI literacy” of their staff and other people operating AI systems on their behalf, taking into account knowledge, experience, education, training and context of use. The article now states it “does not require providers or deployers to guarantee any specific level of AI literacy of any individual” (text on the Commission’s AI Act Service Desk, which flags the article as amended; confirmed 2026-10-05).
- The Commission and Member States must support providers and deployers, with attention to SMEs, the Commission is to publish practical compliance examples, and the AI Board develops recommendations aligned with European competence frameworks. A law-firm analysis (lawandtechnology.eu) points to the AILit Framework (four domains, 22 competences) and DigComp 3.0 as the reference frameworks.
- Earlier claims in this note that the duty was “retained unchanged” or “removed” are superseded: it was retained and softened. Enforcement sits with national authorities (per the Commission Q&A); the Q&A also advises keeping records of training.
State of AI education in 2026
Stanford’s 2026 AI Index (2026-04-13) reports that over 80% of students use AI while only 6% of teachers report clear institutional policies, and that experts and the public differ sharply on AI’s effect on jobs (73% vs 23% expect positive effects). See future-trends-in-ai and ethics-in-ai.
Elements of AI literacy education
(The author’s own summary list, not drawn from a single cited framework.)
- Basics: what AI, machine-learning, neural-networks and large-language-models are and are not.
- Critical thinking: verifying AI output, spotting hallucinations and bias (see massive-multitask-language-understanding; benchmark scores and reliability are different things (author’s view)).
- Ethics and risk: healthcare, finance, law enforcement and employment uses.
- Privacy awareness: what data goes into AI tools and what rights people have - see data-privacy-in-ai.
- Work change: upskilling for tasks complementary to AI.
- Policy literacy: knowing the regulatory landscape.
- Lifelong learning and access: the field changes quickly; programmes should be accessible to all (author’s view).
- Hands-on practice and interdisciplinary framing.
Sources (accessed 2026-10-05)
- https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-4 (current Art. 4, flagged as amended by the Omnibus)
- https://lawandtechnology.eu/en/ai-literacy-digital-omnibus-article-4-ai-act/ (secondary, law firm)
- https://digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers (accessed 2026-09-30)
- https://hai.stanford.edu/ai-index/2026-ai-index-report