European Union: EU reaches landmark deal on AI regulation

1. The EU AI Act has been in development since the EU Commission’s Proposal for a Regulation on AI in 2021.
2. Recent debates have focused on regulating AI foundation models and the use of AI in law enforcement.
3. The Act takes a risk-based approach, categorizing technologies into different risk levels with corresponding obligations and restrictions.
4. Prohibited technologies include biometric identification systems, manipulative techniques, and social scoring systems.
5. The Act also covers foundation models with transparency requirements for all models and stronger requirements for powerful models.
6. Providers of high-risk AI must adhere to transparency and safety constraints to protect health, safety, human rights, and democracy.
7. Penalties for prohibited practices can be up to EUR 35 million or 7% of a company’s annual global revenue.
8. The EU AI Act does not apply outside the scope of EU law and has exceptions for specific provisions.
9. Some technology groups and European companies have expressed concerns about potential innovation stifling.
10. Compliance with the EU AI Act is part of a business’ Responsible AI governance program, with ongoing developments in AI regulation globally.