AI literacy is an obligation now, and almost nobody has the current text

Article 4 of the EU AI Act covers AI literacy. It was rewritten by Regulation (EU) 2026/1744, which came into force on 27 July 2026.

Before that date it required providers and deployers to ensure, to their best extent, a sufficient level of AI literacy among their staff. That was a result you had to prove.

Since that date it requires them to take measures to support the development of AI literacy, and it states explicitly that this does not require guaranteeing any particular level of AI literacy in any individual.

The obligation is still binding. What changed is its shape. It is no longer about proving a standard in each person. It is about being able to show that you did something reasonable and deliberate.

Providers and deployers of AI systems.

Provider is the obvious one: you build or supply the system. Deployer is the one people miss. A deployer is any organisation using an AI system in the course of its work, under its own authority.

That is most companies now. If your team drafts with a language model, screens applicants with a tool, summarises documents, routes support tickets or generates code with an assistant, you are a deployer.

The law does not hand you a curriculum. What it asks for is effort you can evidence, which in practice means training you can describe afterwards.

What was taught. To whom. When. What it covered for the particular systems your people actually use. And whether it was repeated as those systems changed.

A company that can produce that has met a means-oriented obligation. A company that cannot has nothing to show, whatever it believes about its own competence.

There is no certification. There is no accredited trainer status. There is no official curriculum and no register of approved providers.

Nobody can sell you a certificate that makes you compliant, and anybody implying otherwise is selling you something the regulation does not recognise.

We would rather tell you that than let you believe it, because the same honesty is what you are buying when you hire someone to advise you on governance.

Most AI literacy training is a legal briefing with slides, delivered by people who have never deployed the systems they are describing. It is accurate and it changes nothing, because your team leaves knowing what the regulation says and still not knowing what to do differently on Monday.

The useful version starts from what your people actually run. Which tools, on what data, with which decisions attached, and where the human oversight genuinely sits rather than where the policy says it sits.

That requires someone who has shipped software, handled personal data under the GDPR, and written governance documents that had to survive contact with a real system.

Training for in-house teams, in your context and against your tools, with a written record of what was covered so the obligation is evidenced rather than asserted.

One to one sessions for founders, managing directors and the person who has quietly been made responsible for AI without being given time to learn it.

In German or English. Online or in the room. In Vienna, across Austria, and anywhere in Europe by arrangement.

AIPathTech is run by Maria Catalina Kovacs, a software engineer of twenty eight years who has taught programming professionally, including for the public employment service, and who builds AI systems under GDPR Article 9 conditions in her own product.

Article 4 is one obligation among many. If you deploy a high risk system, a great deal more applies to you, and training is not the part that will keep you out of trouble.

We will say so in the first conversation rather than after the invoice.