Europe has achieved a landmark agreement on regulating artificial intelligence (AI) use within the European Union. This deal covers AI in government biometric surveillance and systems like ChatGPT, positioning the EU as a global AI standard-setter.

After intense negotiations and a 24-hour debate, EU countries and Parliament members reached a provisional accord. The agreement requires transparency from foundational AI models and general-purpose AI systems, including technical documentation, copyright compliance, and data summary dissemination.

High-impact models face additional scrutiny, such as risk assessments and cybersecurity measures. Governments can use real-time biometric surveillance in specific cases only. The legislation bans harmful practices like facial image scraping, social scoring, and political inference.

Consumers have the right to file complaints, with fines for violations ranging from €7.5 million to €35 million. While some praised the regulations, business group DigitalEurope criticized them as burdensome, and privacy advocates expressed concerns.

The legislation is set to take effect next year, following ratification, with a two-year grace period.

Europe’s move may influence global AI regulations, offering an alternative to the US’s lighter approach and China’s interim rules.