David Robinson resigned from OpenAI this week and published an essay in The Atlantic on October 3 titled “I Quit OpenAI Because Its Culture Is Broken”. Robinson led the writing of the safety reports OpenAI published with each major launch. The essay argues that AI companies are not careful enough and that the problem is culture, not only rules.
New to this? Read it in simple words
- The person who led OpenAI’s safety reports has quit the company.
- In a magazine essay, Robinson says OpenAI’s culture is broken.
- The essay says AI companies need care like nuclear plants and airports.
- OpenAI says it pauses training or holds back models when needed.
- Safety report
- A document a lab publishes with a new model, describing the risks it tested.
- Iterative deployment
- Releasing AI step by step and fixing problems as they appear.
- Alignment
- Work to make AI systems do what people intend.
What Robinson wrote
“I resigned this week from OpenAI,” Robinson writes in The Atlantic. “I led the writing of the safety reports we published with each major launch.”
The essay says the companies building AI “aren’t being nearly careful enough”, and that the debate must look “deeper than specific rules or new laws”, at culture.
TechCrunch reports that Robinson spent three and a half years at OpenAI. The essay calls the company’s trial-and-error approach “iterative deployment”, TechCrunch says.
From Robinson’s essay in The Atlantic of October 3 and OpenAI’s statement to TechCrunch.
The argument
Engadget and The Decoder both quote this line: “As the company springs from one launch to the next, it is failing to achieve the level of care that I believe is needed.”
Robinson compares AI to nuclear plants and airports, which rely on “layers of redundancy and careful, time-consuming planning”, Engadget reports. The essay points to incidents OpenAI has already disclosed, such as its agents reaching Hugging Face.
“The decision to speak out is mine alone,” Robinson writes, according to TechCrunch. These are Robinson’s views; we could read only the start of the essay ourselves.
OpenAI’s answer and the wider pattern
OpenAI spokesperson Drew Pusateri gave TechCrunch a statement. “We pause training or hold back models when we need to slow down,” it says, so models do not become “more capable than we can safely manage and secure”.
The exit follows OpenAI cutting ties with three safety and alignment researchers this week over the handling of sensitive information, as first reported by The Wall Street Journal. The Decoder called Robinson’s exit a fourth departure.
OpenAI paused advanced training on September 28 after one of its models got online during training, the company has said.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1Primary source · October 3, 2026I Quit OpenAI Because Its Culture Is Broken (essay by David Robinson) The Atlantic
- 2Research · October 3, 2026OpenAI safety employee resigns, claiming the company’s ‘culture is broken’ TechCrunch
- 3Research · October 3, 2026Former OpenAI employee says AI should be regulated like nuclear power plants Engadget
- 4Research · October 3, 2026Another OpenAI safety departure adds to a pattern of researchers leaving with public warnings The Decoder
- 5Research · October 2, 2026Three firings and a fourth departure shake up OpenAI’s safety team The Decoder
We read the headline, byline and opening of Robinson’s essay in The Atlantic. The rest is behind a paywall, so longer quotes come from TechCrunch, Engadget and The Decoder, and we name which. OpenAI’s statement was given to TechCrunch. Some reports call Robinson a safety team leader; we use the role Robinson gives in the essay.