EU AI Act risk class check: which class does your AI fall into?
Four questions on your AI system’s role, area of use and functions. Then you see the likely risk class, your main duties and the dates that apply to you. Your answers stay in your browser.
Not legal advice · dates per Digital Omnibus (EU) 2026/1744
Your result: class, duties, dates.
Answer the four questions above. The likely risk class, the duties for your role, the key dates and the next steps then appear here.
Guide: EU AI Act Article 50 →The EU AI Act explained →
Email me the result
Optional: we send you the result. Otherwise your answers only stay in your browser.
Which risk classes does the EU AI Act define?
The EU AI Act regulates AI by its risk to people. The higher the risk, the stricter the duties.
| Risk class | Examples | Since / from |
|---|---|---|
| Prohibited (Art. 5) | Social scoring, manipulative AI | 2 Feb 2025; two further bans from 2 Dec 2026 |
| High-risk (Annex III and I) | Recruitment, credit checks, medical devices | 2 Dec 2027 / 2 Aug 2028 |
| Transparency (Art. 50) | Chatbots, assistants, AI-generated content | 2 Aug 2026 |
| Minimal risk | Spam filters, translation | no specific duties beyond AI literacy |
General-purpose AI models (GPAI) carry their own duties for their providers since 2 August 2025.
How to determine your AI’s risk class.
Clarify the role
Providers build a system or place it on the market under their name. Deployers use it professionally.
Rule out the bans
Check the prohibited practices under Art. 5: they apply regardless of everything else.
Check the area of use
Annex III areas or a safety component under Annex I usually mean high-risk.
Check the functions
Chat, generated content and emotion recognition trigger the transparency duties under Art. 50.
What usually applies to an AI assistant in a company.
An internal knowledge assistant or a website chatbot usually falls under the transparency duties of Art. 50: people must be able to tell they are talking to an AI. On top come measures for AI literacy under Art. 4, such as an AI policy and a short training. It only becomes high-risk if the system is used in an Annex III area, such as recruitment.
Questions about the risk classes.
Which risk classes does the EU AI Act have?
The EU AI Act distinguishes prohibited practices, high-risk systems, systems with transparency duties and minimal-risk systems. On top come separate duties for providers of general-purpose AI models. The higher the risk to people, the stricter the requirements for providers and deployers.
Is an AI chatbot in a company high-risk?
Usually not. A chatbot for knowledge or customer service usually falls under the transparency duties of Art. 50: users must be able to tell they are talking to an AI. It only becomes high-risk if used in an Annex III area, for example to select job applicants or assess credit.
When do the high-risk duties apply?
Under the Digital Omnibus, Regulation (EU) 2026/1744, the duties for high-risk systems under Annex III apply from 2 December 2027. For AI as a safety component of regulated products under Annex I they apply from 2 August 2028. Bans and transparency duties were not postponed.
Am I a provider or a deployer?
A provider develops an AI system or places it on the market under its own name. A deployer uses it professionally under its own responsibility. Anyone who changes a system substantially or offers it under their own name can become a provider. Clarify your role for each system with your legal counsel.
How reliable is the result?
The check shows the likely classification based on your answers and the law as of 13 October 2026. It does not replace a legal review, especially in borderline cases such as the exception in Art. 6(3). We check deadlines and guidelines monthly and update the check when something changes.
Your AI is classified, now meet the duties.
RAGSuite comes with an AI notice, a source on every answer and an audit log. Try the RAGSuite Engine free for 7 days or talk to us.