Bilancia tra alfabetizzazione all IA e norma europea: articolo 4 AI Act e Digital Omnibus

Immagine creata con IAAI-generated image

AI literacy: the Digital Omnibus rewrites Article 4 of the AI Act

On 24 July 2026 the Official Journal of the European Union published Regulation (EU) 2026/1744 (EUR-Lex), the Digital Omnibus on AI, in force from 27 July. Among its amendments to the AI Act is the full rewriting of Article 4, the provision on AI literacy: the obligation on providers and deployers is not repealed, but its nature changes. Meanwhile the European Commission and the Organisation for Economic Co-operation and Development (OECD) give a shared content to what it actually means to be AI literate. The two levels are best read together.

Article 4 of the AI Act: the literacy obligation until 27 July 2026

In the original version of Regulation (EU) 2024/1689 (EUR-Lex), applicable from 2 February 2025, Article 4 required providers and deployers to take measures to “ensure, to their best extent, a sufficient level of AI literacy” of their staff and of those operating the systems on their behalf, taking into account technical knowledge, experience, education and the context of use. The wording was result-oriented: to secure an adequate level, albeit “to their best extent”.

The Digital Omnibus rewrite: from “ensuring a sufficient level” to “supporting development”

Article 1, point 5, of Regulation (EU) 2026/1744 replaces Article 4 in full. Under the new text, providers and deployers “shall take measures to support the development of AI literacy” of their staff and of those operating the systems on their behalf, still taking account of skills, experience, education and context. The legislator adds a clarification that shifts the axis of the obligation: it “does not require providers or deployers to guarantee any specific level of AI literacy of any individual”. A second paragraph tasks the Commission and the Member States with supporting and facilitating the efforts of providers and deployers, with particular regard to small and medium-sized enterprises. The move, consistent with the omnibus’s simplification logic that we examined when commenting on its publication in the Official Journal, turns an obligation of result into an obligation of means.

The content of literacy: the AILit Framework and DigComp 3.0

If the rule no longer sets a threshold, it remains to be established what AI literacy actually is. A now-consolidated European framework answers that question. “Empowering Learners for the Age of AI” (AILit), a joint initiative of the European Commission and the OECD with the support of Code.org, presented for consultation in May 2025 and finalised in 2026, structures literacy for primary and secondary education around four domains (Engage with AI, Create with AI, Manage AI, Design AI) and 22 competences, and will feed the PISA 2029 assessment. On citizens’ digital competence, the fifth edition of the European framework, DigComp 3.0 (Joint Research Centre, JRC, 2025), integrates AI competence transversally and, for the first time, introduces over 500 learning outcomes. Alongside these sit the United Nations Educational, Scientific and Cultural Organization (UNESCO) frameworks on AI competencies for students and teachers. None of these instruments is binding, yet together they form the substantive reference for the “what”, where the rule now confines itself to the “how”.

Critical points: a lighter obligation and the compass of voluntary frameworks

The combination deserves attention. On one hand, the reform reduces the compliance burden: the absence of a guaranteed level lessens the risk of disputes over “how much” a given individual must know, but leaves the actual measure of the effort to the operator’s diligence. On the other, just as the law retreats on the result, pedagogy advances on the content, offering verifiable criteria of what literacy should encompass. We consider that the centre of gravity thus shifts from formal compliance to the substantive quality of training: for providers and deployers the obligation remains as one of means; for the Commission and Member States it becomes a task of promotion, with an eye to small and medium-sized enterprises. The theme sits within the broader picture of European digital competences, which we set out when commenting on the State of the Digital Decade 2026.

Conclusions

What remains of Article 4 is its backbone, not the obligation of result: providers and deployers must support AI literacy, without having to guarantee an individual level. The “what” is now entrusted to the European frameworks, from AILit to DigComp 3.0, which become the voluntary compass for anyone seeking to give substance to an obligation the law has made lighter.

Author: Valentina Grazia Sapuppo


AI AnthropoCosmic In evidenzaAI AnthropoCosmicCall for Paper aperta fino al 15 settembre 2026. Un progetto internazionale per un’IA a servizio dell’Uomo, dell’Ambiente e del Cosmo. Leggi l’articoloAI Open Mind AI AnthropoCosmic FeaturedAI AnthropoCosmicCall for Paper open until 15 September 2026. An international project for an AI at the service of humanity, the environment and the cosmos. Read the articleAI Open Mind Agentic AI In evidenzaAgentic AILimiti prima dell’azione, evidenze durante, responsabilità dopo. Il volume di Nicola Fabiano sulla governance dei sistemi agentici, con la prefazione di Antonino Caffo.Capitolo 16 a cura dell’Avv. Valentina Grazia SapuppoLeggi l’articolo Agentic AI FeaturedAgentic AILimits before the action, evidence during, responsibility afterwards. Nicola Fabiano’s book on the governance of agentic systems, with a preface by Antonino Caffo.Chapter 16 by Valentina Grazia SapuppoRead the article