What are the EU AI Act risk tiers?

The EU AI Act sorts AI systems into four tiers: unacceptable risk (prohibited practices), high-risk (heavily regulated uses listed in the Act), limited-risk (transparency obligations, such as disclosing that users are talking to AI or that content is AI-generated), and minimal-risk (no specific obligations). The rules phase in through 2026 and beyond.

Does the EU AI Act apply to a company outside the EU?

It can. The Act applies to providers and deployers whose AI system output is used in the EU, regardless of where the company is based. If you serve EU users or your system's output reaches the EU market, the obligations can attach to you.

Is this classification legal advice?

No. This is an educational tool that gives a directional classification based on your answers. The Act's definitions, annexes, and exceptions are detailed and fact-specific. Use this as a starting point, then get a proper readiness review before relying on it.

Yes, it is free with no signup. It exists to help you understand where your AI system likely sits before you plan a compliance program.

EU AI Act Risk Classifier | Free Tool

Step 1 · Prohibited practices (Article 5)

Does your AI system do any of these?

These practices are banned outright under the EU AI Act. Tick any that apply.

If none apply, leave them all unchecked and continue.

Step 2 · High-risk uses (Annex I & III)

Is your system used in any of these contexts?

These are the regulated high-risk domains. Tick any where your system is used to make or materially influence decisions about people.

Only tick a box if the AI meaningfully affects the outcome, not if it is a trivial helper.

Step 3 · Transparency triggers (Article 50)

Does your system do any of these?

These trigger limited-risk transparency duties even when the system is not high-risk.

Obligations that typically follow

How this works, and why it is not legal advice. This tool applies the EU AI Act's own tiering logic: prohibited practices are checked first, then high-risk uses, then transparency triggers, and anything left is minimal-risk. Classification in the real Act depends on precise definitions, annex entries, and exceptions that are fact-specific. Obligations and deadlines phase in on a staggered timeline through 2026 and beyond. Treat this as a directional starting point, not a compliance determination.

Not ready for a call yet?

Get the EU AI Act playbook

A few short notes from Jacob on getting an AI product ready for the EU AI Act without drowning in the annexes. No fluff, unsubscribe in one click. Reply anytime; it reaches him directly.

From Jacob Masse, founder of traztech. No spam, unsubscribe in one click.

Want it done for you?

EU AI Act Readiness

We map your AI systems to the Act, confirm your risk tier, and build the compliance evidence you will need.

Explore EU AI Act Readiness →

EU AI Act Risk Classifier

Does your AI system do any of these?

Is your system used in any of these contexts?

Does your system do any of these?

Obligations that typically follow

Get the EU AI Act playbook

Need to get EU AI Act ready?