Ex-OpenAI Policy Researcher Slams AI Firms for Weakening Safety Commitments
Miles_Brundage · x · 2026-08-04
Former OpenAI policy researcher Miles Brundage criticized frontier AI companies for backsliding on their safety commitments. He noted that these companies initially pledged to avoid causing massive harm, but later deemed those standards too strict, adopting weaker policies only when mandated by new California regulations.
He further pointed out that the US lacks a substantive quality bar for AI safety policies beyond topic coverage and detail. This low threshold, he argues, ironically frames those advocating for stronger safeguards as enemies of the industry.
Related event: Ex-OpenAI Advisor Slams AI Firms' Retreating Safety Policies(4 posts)→
More from AGI Musings
- Robots Handling Errands Could Shift 50M Gig Workers to Dexterity Labor — willccbb · 2026-08-04
- Karpathy Interview: AGI is a Decade Away, Current RL is Terrible — jiqizhixin · 2026-08-04
- Humanoid Robot Production Could Reach 50M in 5 Years, Driven by Scaling Effects — willccbb · 2026-08-04
- Expert Predicts Longevity Escape Velocity is Unlikely, Sets Median Lifespan Goal of 110 — MaxUnfried · 2026-08-04
- AMD Faces AI Earnings Test as Open Models Threaten Closed Labs' Margins — AccBalanced · 2026-08-04
- OpenAI Alignment Team Under 2% of Staff, Ex-Researchers Urge Pivot to Safety — Miles_Brundage · 2026-08-04