Article 4, explained
What the EU AI Act requires of your company, without the jargon and with the legal reference in plain sight.
Which AI systems in HR count as high-risk under Annex III
Screening CVs, assessing performance or helping decide a promotion with AI is not just any use: Annex III classifies it as high-risk. What that actually means and what obligations it adds.
What not to type into ChatGPT (or any AI) at work
The most common risk under Article 4 is not failing to use the tool: it is using it well without knowing what not to feed it. A concrete guide to what data must never leave your company for a public AI.
Who does Article 4 of the EU AI Act apply to?
Article 4 has been binding since February 2025, and most companies do not know whether it covers them. Who is in scope, who is not, and what «sufficient AI literacy» actually means.
Provider or deployer: how to tell which one you are
Every other obligation hangs off this distinction, and it is the one most companies get wrong. Four concrete situations, and the case almost nobody sees coming: being both at once.
What penalties apply for breaching Article 4?
The 35 million figure circulates everywhere and is almost always misapplied. What fine actually attaches to a lack of training, who imposes it, and what happens long before you get there.
What an inspection asks for to accept your AI training
Having trained your staff and being able to prove it are two different things, and only the second counts before an authority. What constitutes evidence, what does not, and how long to keep it.