Blog

Is your AI system high-risk? The Annex III categories teams should screen first

A practical list of Annex III-style categories and how teams should use them as a pre-evaluation framework, not as a replacement for legal review.

Published: ยท Updated:

Use Annex III as a pre-evaluation lens, not as a final verdict

Teams usually need a fast way to decide whether documentation work should start now. Annex III-style categories give you that pre-evaluation lens. Recruitment and worker management, educational access and assessment, access to essential services such as credit, and parts of healthcare support are common early flags.

Why customer support systems are trickier than they look

A customer support agent is not automatically high-risk. But the moment it influences decisions about eligibility, financial outcomes, or regulated access paths, the system becomes more serious. That is why a builder should ask both what the system is called and what it actually does in production.

What to do after a preliminary high-risk result

Do not wait for perfect certainty. Build the package structure, document assumptions, start technical evidence collection, and then validate the legal scope. That is a much cheaper path than doing nothing and discovering late that the evidence layer is missing.

Next step

Continue with Builder or the starter

FAQ

Frequently asked questions

How should a team use the high-risk AI list?

Use the list as a scoping input, not a final answer. The package still needs role, intended purpose, Annex III or product-safety context, evidence links, and human review.

How does "Is your AI system high-risk? The Annex III categories teams should screen first" connect to the product?

The article explains the buyer and compliance context. The product pages then turn that context into a package path, templates, starter commands, and evidence handoff steps.

Does the article replace legal review?

No. The article is informational context. Legal and compliance teams still need to review the selected package, role, scope, article coverage, and final evidence sufficiency.