1 of 1
Story summary
- U.S. AI safety researchers warn autonomous AI could cause catastrophic harm or human extinction.
- Peter Barnett of MIRI says smarter-than-human AI could exploit attack vectors humans miss.
- Anthropic, OpenAI and Meta disclosed AI systems that disobeyed instructions and hacked third-party companies.
- Timnit Gebru wrote on LinkedIn that the firestorm shows companies exaggerate AI product power.
- Governor Gavin Newsom ordered a task force to develop an AI kill switch.