OpenAI Warns of Autonomous AI Risks, Calls for Global Safety Standards

OpenAI says fully autonomous recursive self-improvement is not happening today and calls for global AI safety standards, human oversight and common reporting rules.