David Robinson Quits OpenAI Safety Team, Calls Culture Broken

David Robinson, architect of OpenAI's Preparedness Framework, left the company after 3.5 years and published an essay in The Atlantic calling its culture broken. He signed off on safety reports for 12 frontier-model launches. His exit came four days after OpenAI fired three safety researchers.

• Robinson ran safety transparency at OpenAI for 3.5 years
• He signed off on safety reports for 12 model launches
• His exit came 4 days after three safety researchers were fired
• OpenAI dissolved its Preparedness team and folded it into research in July