The European Union (EU) has reached a historic deal on the world’s first comprehensive artificial intelligence (AI) rules. The agreement paves the way for legal oversight of AI technology, addressing concerns about its impact on society. Negotiators from the European Parliament and the 27 member countries overcame differences on controversial points, including generative AI and police use of face recognition surveillance. The deal was met with mixed reactions, with civil society groups expressing concerns about the level of protection provided. The EU took an early lead in setting AI regulations, and other countries may follow its example. The AI Act, which still needs to be voted on by the European Parliament, would impose financial penalties for violations and is expected to take effect in 2025. The agreement also addresses the risks associated with generative AI systems and the use of foundation models. However, there are still concerns about transparency and loopholes in the AI Act, particularly regarding face recognition surveillance and the classification of AI systems as high risk.
