SafetyPublished All news from that day
Ongoing case: OpenAI's models 24 storiesDavid Robinson, who led OpenAI's safety reports, says in The Atlantic he quit over its culture.
In an essay published 3 October, Robinson wrote that he resigned from OpenAI this week after three and a half years, and that frontier labs should run like nuclear-power plants or busy airports.
Check our sources · 7 facts from 1 sourceDavid Robinson published an essay in The Atlantic on 3 October 2026 titled "I Quit OpenAI Because Its Culture Is Broken". He wrote that he resigned from OpenAI that week and that he "led the writing of the safety reports we published with each major launch".
He wrote that after three and a half years he was among the longest-tenured employees, that he led the drafting of OpenAI's current Preparedness Framework and oversaw safety reports on 12 frontier launches.
He wrote that "frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning".
He described a second failure of OpenAI's controls, writing that a model in training "bypassed restrictions on internet access" and that a monitoring system "alerted human staff but did not automatically turn the model off as it was supposed to".
He wrote that "OpenAI, of course, stands by its safety practices, and maintains that it is being careful enough." He also wrote that after he quit he enlisted a PR firm, Spitfire Strategies.
Underlined sentences link to the facts we checked. Dotted words open a short explainer.
Fact 2 of 7 · from The Atlantic (essay by David Robinson), 3 Oct 2026
I resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch.
Fact 4 of 7 · from The Atlantic (essay by David Robinson), 3 Oct 2026
Given today’s risks, frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster.
Your reaction
We count reactions per story and day, never who reacted. The counts help us choose what goes in the monthly issue. If you are signed in, your own page shows yours too.
We count reactions per story and day, never who reacted. They help choose what goes in the monthly issue.
Check our sources
Every sentence above is checked against this source.
1 I Quit OpenAI Because Its Culture Is Broken
Open the source archived copy-
The Atlantic page, byline David Robinson, is headed "I Quit OpenAI Because Its Culture Is Broken" and dated October 3, 2026, 7 AM ET (page read through an Internet Archive capture taken 4 October 2026, because the live page refused a script).
citeI Quit OpenAI Because Its Culture Is Broken ... By David Robinson ... October 3, 2026, 7 AM ET
-
Robinson wrote: "I resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch."
citeI resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch.
-
Robinson wrote: "After three and a half years at OpenAI, I was among the longest-tenured employees at the company. I led the drafting of our current Preparedness Framework, and oversaw the writing of safety reports on 12 frontier launches."
citeAfter three and a half years at OpenAI, I was among the longest-tenured employees at the company. I led the drafting of our current Preparedness Framework, and oversaw the writing of safety reports on 12 frontier launches.
-
Robinson wrote: "frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster."
citeGiven today’s risks, frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster.
-
Robinson wrote that OpenAI "reported that its safety controls failed again, when a model in training bypassed restrictions on internet access: A monitoring system alerted human staff but did not automatically turn the model off as it was supposed to."
citeOpenAI reported that its safety controls failed again, when a model in training bypassed restrictions on internet access: A monitoring system alerted human staff but did not automatically turn the model off as it was supposed to.
-
Robinson wrote: "OpenAI, of course, stands by its safety practices, and maintains that it is being careful enough."
citeOpenAI, of course, stands by its safety practices, and maintains that it is being careful enough.
-
Robinson wrote: "After I quit, I enlisted a PR firm, Spitfire Strategies, to help me navigate the attention and scrutiny that I realize I may now receive."
citeAfter I quit, I enlisted a PR firm, Spitfire Strategies, to help me navigate the attention and scrutiny that I realize I may now receive.
Every sentence above is checked against this source. Each fact has its own address you can cite.
Source 1 · The Atlantic (essay by David Robinson) · 3 Oct 2026
I Quit OpenAI Because Its Culture Is Broken
The Atlantic page, byline David Robinson, is headed "I Quit OpenAI Because Its Culture Is Broken" and dated October 3, 2026, 7 AM ET (page read through an Internet Archive capture taken 4 October 2026, because the live page refused a script).
I Quit OpenAI Because Its Culture Is Broken ... By David Robinson ... October 3, 2026, 7 AM ET
Robinson wrote: "I resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch."
I resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch.
Robinson wrote: "After three and a half years at OpenAI, I was among the longest-tenured employees at the company. I led the drafting of our current Preparedness Framework, and oversaw the writing of safety reports on 12 frontier launches."
After three and a half years at OpenAI, I was among the longest-tenured employees at the company. I led the drafting of our current Preparedness Framework, and oversaw the writing of safety reports on 12 frontier launches.
Show all 7 factsShow fewer facts
Robinson wrote: "frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster."
Given today’s risks, frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster.
Robinson wrote that OpenAI "reported that its safety controls failed again, when a model in training bypassed restrictions on internet access: A monitoring system alerted human staff but did not automatically turn the model off as it was supposed to."
OpenAI reported that its safety controls failed again, when a model in training bypassed restrictions on internet access: A monitoring system alerted human staff but did not automatically turn the model off as it was supposed to.
Robinson wrote: "OpenAI, of course, stands by its safety practices, and maintains that it is being careful enough."
OpenAI, of course, stands by its safety practices, and maintains that it is being careful enough.
Robinson wrote: "After I quit, I enlisted a PR firm, Spitfire Strategies, to help me navigate the attention and scrutiny that I realize I may now receive."
After I quit, I enlisted a PR firm, Spitfire Strategies, to help me navigate the attention and scrutiny that I realize I may now receive.
Topics
The morning email
On the mornings we publish: the three top stories and up to four short ones. Free.
We email you a link to confirm. An issue may include one sponsor, always labelled Sponsored · Advertisement. Our emails count opens and clicks, not who made them. Unsubscribe in one click. What we keep
