OpenAI Safety Report Author Quits, Warns of 'Broken' AI Culture — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- David Robinson resigned from OpenAI this week and published an editorial in The Atlantic, declaring the AI industry's culture "broken" in ways that go beyond new rules or regulations.
- Robinson previously wrote the safety reports that accompanied every major model release at OpenAI, and now accuses Silicon Valley of running on "extreme confidence," "perpetual sprints," and "unimpeded optimism."
- Robinson argues frontier AI labs should operate like nuclear power plants or busy airports, with "layers of redundancy and careful, time-consuming planning" so that inevitable human error doesn't open a door to disaster.
- Robinson joins a growing exodus of publicly departed safety researchers that includes Jacob Coxon (Anthropic), Robert O'Callahan, Bilal Chughtai, and Josh Engels (Google DeepMind), and Joe Benton (Anthropic).
- Jacob Coxon, who appears to have kicked off the wave by quitting Anthropic, publicly warned that AI "could kill us all by the end of the decade."
Why it matters: The people who wrote OpenAI's safety reports are now saying the culture around those reports is hollow — and at least six safety researchers from OpenAI, Anthropic, and Google DeepMind have publicly departed, a pattern that undermines internal safety claims at exactly the moment regulators are weighing how much to trust self-reporting by frontier labs.
Ask SkimNews
