The AI uses the Act calls high-risk.
If one of your systems is used for something on this list, it is likely high-risk, and the full set of deployer obligations applies. Here is the list, broken into plain use cases.
This content is a development placeholder, not reviewed by counsel. The real Annex III ruleset and document text are authored and reviewed by qualified legal counsel before they are shown as authoritative. Not legal advice.
Prohibited
2 usesBanned outright under Article 5. If a system does one of these, the answer is to stop, not to document.
High-risk (Annex III)
12 usesAllowed, but these carry the heaviest deployer obligations: risk management, human oversight, technical documentation, logging, and more.
Limited-risk (Article 50)
3 usesMostly a duty to be transparent, like telling people they are dealing with AI or labelling synthetic content.
Does this describe one of your systems?
Run the free diagnosis to see which of your tools fall under Annex III, and what you would owe if they do. No signup.
14-day trial of Business · no card · see pricing