AI Safety Newsletter AISN #59: EU Publishes General-Purpose AI Code of Practice
Jul 15, 2025
Discover the EU's groundbreaking General-Purpose AI Code of Practice, which aims to enhance safety and transparency in AI systems. Learn about the significant restrictions on AI uses, including bans on social scoring and predictive policing. Dive into Meta's ambitious revamp of its superintelligence initiatives and their strategies to attract top talent from competitors. Explore the latest in AI legislation and industry controversies, underscoring the urgent need for effective risk management and governance in the evolving AI landscape.
AI Snips
Chapters
Transcript
Episode notes
EU AI Code Enhances Clarity
- The EU's AI Act bans certain AI uses and limits others deemed high risk, focusing heavily on transparency and safety for general-purpose AI (GPAI).
- The General Purpose AI Code of Practice clarifies safety and systemic risk requirements ahead of the Act's enforcement on August 2, 2025.
Implement Rigorous Risk Frameworks
- GPAI providers must implement structured frameworks to identify, analyze, and mitigate systemic risks like CBRN hazards and harmful manipulation.
- They must continuously monitor deployed models, report incidents promptly, and adapt safety methods as AI capabilities advance.
EU Code Sets Industry De Facto Standard
- OpenAI and Mistral have indicated intent to comply with the EU Code of Practice, signaling its influence in shaping industry safety standards.
- With frontier AI companies likely complying, similar U.S. legislation may become less urgent for AI safety.
