EU AI Act sets risk-based rules for AI systems

The AI Act would create a legal framework for selling and using AI in the EU, with consistent standards across Member States. Its objectives include protecting health, safety and fundamental rights while ensuring AI systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly.

The framework applies to systems placed on the market, put into service or used in the EU, including offerings from global providers that make systems or outputs available to EU users. It divides AI into risk categories: unacceptable-risk systems would be prohibited, high-risk systems would require assessments and registration, limited-risk systems would face transparency duties, and minimal-risk systems would carry no mandatory obligations but may follow voluntary codes of conduct.

Prohibited practices could trigger fines of up to EUR 40 million or up to 7% of a company’s global annual turnover, whichever is higher, though penalties are expected to be proportionate. The regulation is likely to be adopted in early 2024, followed by a transitional period of at least 18 months before full enforcement.