OpenAI Proposes Global AI Safety Standards

OpenAI released international AI safety standards on Monday, targeting misalignment risks and calling for cooperation on frontier AI safety. The proposals recommend building on existing safety institutes and warn against pursuing fully autonomous recursive self-improvement unless done safely.