The European Commission Says AI Literacy Stays an Obligation With No Specific Level Mandated

The European Commission's AI Office keeps a page of questions and answers on AI literacy. It explains Article 4 of the EU's AI Act. The page was last updated on October 20, 2026.

Article 4 requires providers and deployers of AI systems to take measures to support the AI literacy of their staff. It also reaches other persons dealing with those systems on their behalf. These are not employees, the page says, but people broadly under the organization's remit. Its examples are a contractor, a service provider and a client. Parts of the page follow, not the whole.

Are there requirements for specific industries, including financial services and healthcare? No, the AI Office says. It does not impose requirements for specific sectors. Yet context should be relevant when an organization develops its approach, the page says. That includes the sector and the purpose of a system. The level of risk of the AI systems should be considered too.

The Proposal and the Amendment

The Commission presented the Digital Omnibus on AI on November 19, 2025. On Article 4, the page says, it proposed a shift. The obligation to promote AI literacy, upskilling and reskilling would move onto the Member States and the Commission. That was in place of enforcing what the page calls "an unspecific obligation on organisations."

The amendments to Article 4 entered into force in mid-July 2026, the page says. Under them, AI literacy remains an obligation for providers and deployers of AI systems. But no specific or sufficient level is mandated. The Commission and the Member States are to continue supporting those efforts. That includes publishing practical examples of how to comply on the Single Information Platform. The AI Board is to adopt recommendations on Article 4, setting out common objectives.

For those who deploy high-risk AI systems, the page says, the obligation to ensure their staff are trained to ensure human oversight "remains in place."

Knowledge, Records, Governance and Training

Article 4 does not entail an obligation to measure employees' knowledge of AI, the page says. Nor does it require providers or deployers to guarantee any specific level of AI literacy of any individual.

Do organizations need certificates to document what they do? There is no need for a certificate, the page says. Organizations can keep an internal record of trainings or other guiding initiatives. Is an AI officer needed, or a governance board? No specific governance structure is mandated to comply with Article 4, it says.

No set level, no duty to measure and no certificate.

Is training mandatory? That depends on the organization's answers to the page's own considerations, it says. It may also depend on the further recommendations the AI Board should prepare. Yet in many cases, it says, simply relying on a system's instructions for use "might be ineffective." So might asking staff to read them.

Elsewhere the page says no strict requirements or mandatory trainings are imposed. The AI Office keeps a living repository of AI literacy practices from companies and the public sector. At the moment, the page says, replicating those practices does not automatically grant a presumption of compliance with Article 4.

Article 26 obliges deployers of high-risk systems to ensure that the staff dealing with them in practice are sufficiently trained. Those staff must be able to handle the system and ensure human oversight. Relying on the instructions for use is therefore not sufficient, the page says. Further measures are necessary.

Employees, Developers and Contractors

The page takes the case of a company whose employees use an AI chatbot to write advertisement text or to translate text. Does it need to comply with Article 4? Yes, the page says. They should be informed about the specific risks, the page says, for example hallucination.

Can people with a degree or experience in AI development be considered AI literate without further action? Normally yes, the page says. But it depends on the AI tool in question and their specific qualification. The organization should still ask whether they know what needs to be known about its AI systems and how to deal with them. It should ask whether they are aware of all the risks and how to avoid or mitigate them. And it should consider what else they might need, such as legal and ethical aspects of AI.

Must a service provider using AI have a contractual obligation to demonstrate AI literacy? That depends on the type of system and its risk, the page says. In general, people working for a service provider or contractor need the appropriate AI skills for the task. So do employees.

Application and Enforcement

Article 4 entered into application on February 2, 2025, the page says. So the obligation already applies. The rules on supervision and enforcement apply from August 2, 2026. National market surveillance authorities supervise and enforce Article 4. The AI Office does not. The national authorities could impose penalties and other enforcement measures, based on national laws.

Any sanction must be proportionate and based on the individual case, the page says. It must take account of factors such as the nature and gravity of the infringement. Whether it was intentional or negligent counts as well. A sanction might be more likely, the page adds, if there is proof of an incident due to a lack of appropriate training and guidance of employees or other persons.