What does the AI Act, the European regulation on artificial intelligence, say?
The AI Act, EU Regulation 2024/1689, is the world's first comprehensive legal framework on AI. In force since August 1, 2024, it classifies AI systems by risk level: prohibited practices, high risk with strict oversight, limited risk subject to transparency, minimal risk free. Application is phased through 2027.
The AI Act, Regulation EU 2024/1689, is the world's first comprehensive legal framework dedicated to artificial intelligence. It entered into force on August 1, 2024, and is based on a simple logic: classify AI systems according to their risk level and impose proportionate obligations. Four levels structure the text: prohibited practices, heavily regulated high-risk systems, limited-risk systems subject to transparency obligations, and minimal-risk systems left free. Its application is gradual, from 2025 to 2027.
The risk-based logic
The more an AI use can affect fundamental rights, safety or health, the more it is regulated. An AI that recommends music is minimal risk. An AI that sorts applications or evaluates credit is high risk. Certain practices are deemed unacceptable and simply prohibited.
The four levels
Unacceptable risk: prohibited practices, such as generalized social scoring or certain forms of manipulation and biometric surveillance.
High risk: uses in recruitment, education, credit, justice, essential public services, biometrics, critical infrastructure. Strict obligations for risk management, data quality, human oversight, transparency, documentation.
Limited risk: transparency obligations, for example flagging that content is artificial or that one is speaking to an AI.
Minimal risk: no specific obligation.
Who it applies to
The regulation targets providers who develop systems and deployers who use them, including businesses, administrations and local authorities. It also applies to actors outside the European Union as long as their systems are used in the Union.
Why it concerns you
It establishes concrete rights: prohibition of abusive practices, transparency requirements, guarantees regarding high-risk systems that affect you. It also creates, as of February 2025, an obligation of AI literacy for organizations that deploy it.
Frequently Asked Questions
When does it fully apply?
In stages until 2027. The prohibitions and AI literacy have applied since February 2025, most obligations from August 2026.
Does it concern individuals?
It does not impose obligations on individual users, but it creates protections from which they benefit.
Where to read the text?
Regulation EU 2024/1689 is available on EUR-Lex.
ICIA Resource
ICIA offers MentivisOS ICIA, a free lifelong platform to learn AI with a personalized learning path.
Discover MentivisOS ICIA