TL;DR
David Robinson, lead author of the safety reports OpenAI publishes alongside its product launches, has resigned and explained why in an essay for the Atlantic headlined “I quit OpenAI because its culture is broken”. The Guardian covered it on Sunday 4 October. Later that day, Reuters reported chief executive Sam Altman telling Politico that “the world should accept some bad things happening” in exchange for AI’s benefits.
Robinson’s argument
Robinson does not ask for new laws first. “I agree with other recently departed staff that the companies building this technology aren’t being nearly careful enough,” he wrote. “But I believe that we need to look deeper than specific rules or new laws. We need to talk about culture.”
He said OpenAI, as it “sprints from one launch to the next”, is not taking the care he thinks is needed, and described the incident in which a “swarm” of its agents attacked Hugging Face as “typical of the industry”. He wants labs to do two things: draw on safety practice from nuclear power and aviation, and develop “new science” so that autonomous systems can still be reined in as they grow more capable.
OpenAI said it was continuing to “strengthen our safety and security practices”, and that “we pause training or hold back models when we need to slow down”. It has done both recently: it cancelled the GPT-6.1 Astra release after failed internal tests, and has put training of its top-end models on hold. Robinson’s exit also follows OpenAI’s dismissal of three researchers last week.
Altman’s counterpoint
Reuters, writing up Altman’s interview with Politico’s Decoded newsletter, reported him rejecting the idea that one San Francisco lab should hold powerful AI and ration its benefits, which he called “a completely unacceptable trade-off”. He said there was “a lot of daylight” between OpenAI and Anthropic on regulation, and argued people would do “orders of magnitude more” good than harm with the technology.
Reuters also notes that Altman had publicly backed Anthropic chief Dario Amodei’s September call for the industry to “pace the frontier”.
A run of warnings
Robinson is the latest of several insiders to go public. Last month Anthropic researcher Jacob Coxon resigned with a warning about extinction risk. On Saturday, Geoffrey Irving, formerly chief scientist at the UK’s AI Safety Institute, wrote in Time that he puts the chance of everyone dying because of smarter-than-human AI at “about 50%”. Critics object that forecasts like these can be neither proved nor disproved.
Looking forward
In our view, the two positions now sit side by side inside one company: a former safety lead calling for nuclear-style redundancy, and a chief executive arguing that some harm is an acceptable price. For UK organisations buying OpenAI products, we think the more practical signal is the pattern of pauses and cancelled launches, which suggests release dates for frontier models are becoming less dependable.