Negotiators from the European Parliament and Council reached a provisional agreement on Friday, Dec 8, regarding the rules governing artificial intelligence (AI) use. This milestone addresses various aspects of AI implementation, ranging from biometric surveillance to regulating AI systems like ChatGPT.

Comprehensive Coverage of Governmental AI Use and Regulation

The agreement extensively covers the governmental use of AI, particularly in biometric surveillance. It also outlines regulations for AI systems, including models like ChatGPT. Additionally, transparency rules are highlighted, emphasizing the necessary steps before AI systems enter the market. This involves technical documentation, adherence to European Union copyright, and the sharing of training content summaries.

EU Takes Lead in AI Legislation with Groundbreaking Agreement

Positioning itself as a pioneer in AI legislation, the European Union (EU) aims to establish the first supranational authority with comprehensive laws on AI usage. The focus is on ensuring the beneficial use of AI while implementing safeguards against potential risks. The agreement results from a substantial 24-hour debate on Dec. 8, followed by 15 hours of negotiations, marking a significant step forward in shaping the regulatory landscape for AI in the EU.

Key Provisions in the Agreement

The agreement outlines specific provisions for AI models with significant impact and systemic risks. These models are required to conduct evaluations addressing potential risks, undergo adversarial testing to ensure system resilience, report incidents to the European Commission, prioritize cybersecurity, and disclose information regarding energy efficiency. These provisions enhance accountability, transparency, and safety in deploying impactful AI systems.

Reference: EU AI Act: first regulation on artificial intelligence | News | European Parliament (europa.eu)