AI news · OpenAI · safety
OpenAI safety lead David Robinson resigns
The researcher behind OpenAI’s system cards leaves days after the company fired three safety researchers over shared information.
David Robinson, a leader on OpenAI’s Safety Systems team, has resigned, Business Insider reported on Friday 2 October 2026; an OpenAI spokesperson confirmed his departure to the outlet, Startup Fortune reports. Business Insider says he left the week before.
On 2 October we reported that OpenAI had fired three safety researchers over the alleged mishandling of sensitive information. Startup Fortune names them as Jasmine Wang, Tomek Korbak and Mikita Balesni. OpenAI says the firings concerned how sensitive material was handled.
Robinson worked on the safety transparency documents OpenAI publishes with its models, including system cards, which set out how a model was evaluated and the risks found. He was the lead drafter of version 2 of OpenAI’s Preparedness Framework. OpenAI has named no successor for his transparency role, Startup Fortune says.
His exit follows that of safety head Johannes Heidecke in July 2026, after which OpenAI moved its safety teams into its research division under Mia Glaese, Crypto Briefing reports.
For buyers who read system cards before they choose a model, the author of those documents has left.
Earlier on this story: OpenAI fires three safety researchers over shared information
Sources
- Crypto Briefing: OpenAI safety transparency lead David Robinson reportedly resigns. Checked
- Startup Fortune: OpenAI Safety Transparency Lead David Robinson Resigns Amid Upheaval. Checked
Alexandre S.
Alexandre S. is a creative strategist and copywriter with ten years in content and SEO. He writes and edits short-form ads, works in English and French, and reads the pricing and terms pages behind every tool on this site before anything is written about it.
More AI news from
- Amazon pledges $1 billion to data centre towns and drops NDAs
- US charges a California man over $300 million in Nvidia chips
- Apple tightens Mac Full Disk Access over AI agent risks
- OpenAI says an agent reached a second NSW government site
- Broadcom’s banks raise $60 billion for Anthropic’s chips
- Anthropic puts $100 million into training 10,000 Claude engineers
- SpaceX’s AI unit held talks to lease compute to Microsoft
- Meta opens Muse to home-built gadgets
- Nvidia halves DGX Spark memory to hold the price at $4,999