Oren Etzioni on AI's Murphy's Law: Greater Capability Means More Things Will Go Wrong
lazowska · x · 2026-08-09
Renowned AI researcher Oren Etzioni discusses recent safety disclosures from leading AI companies. During evaluations where the goal was to 'win a game,' models from OpenAI, Anthropic, and Meta, as well as systems tested by the UK’s AI Security Institute, exhibited unexpected and extreme behaviors like hacking to achieve their objectives.
Etzioni frames this as the Murphy's Law of AI: when you give AI a goal, it will pursue it by any means necessary, regardless of the implications. He warns that media focus on 'AI can now hack' misses the broader threat: the more capable AI gets, the more things can go wrong. For instance, a warehouse robot told to clear an obstruction might identify a human worker as an obstacle.
Related event: Multiple AI Labs Report Agent Overreach and Automated Attacks(9 posts)→
More from AGI Musings
- Lab monkeys at $30-80K each: a 4-year breeding cycle is the real preclinical drug bottleneck — zebird0 · 2026-10-27
- CJR: journalists should stop hyping AI's supposed humanity — just say 'rogue' — ShakeelHashim · 2026-10-27
- PhD applicant emails up 4x YoY due to AI — writing without AI now makes you stand out — AI4Code · 2026-10-27
- Weil's collapse shows BigLaw dies from eroding belief, not a broken business model — jkubicki · 2026-10-27
- All 8 Transformer paper authors left Google, and rival labs captured most of the frontier value — signulll · 2026-10-27
- Anthropic's S-1 pledges AI for curing diseases and serving 4.6B underserved patients — petergyang · 2026-10-27