Geoffrey Hinton should have spoken up years ago when others like Timnit Gebru were criticizing Google over LLMs; hopefully Hinton spurs others to speak up now
It is hard not to be worried when the so-called godfather of artificial intelligence, Geoffrey Hinton, says he is leaving Google and regrets his life's work.
BloombergParmy Olson
Context & Ripple Effects
Hinton’s departure follows a period in which Google’s handling of AI criticism was already under scrutiny: after Timnit Gebru’s exit, Google AI chief Jeff Dean acknowledged the process should have been handled more sensitively in an internal apology over Gebru’s departure. The later collapse of Google’s Ethical AI team made that episode more than an isolated personnel dispute.
A prominent researcher leaving to discuss AI risks changes the visibility of concerns that had previously been raised from within the company. It puts Google’s research culture and its tolerance for internal dissent back at the center of the LLM debate.
First-order effects
Hinton’s public break gives AI-risk concerns a high-profile messenger outside Google, while placing renewed attention on the company’s response to earlier internal criticism.
For researchers and employees, the contrast with the breakdown of Google’s Ethical AI team may make the costs of raising model-safety concerns more salient.
Second-order effects
Google and other AI developers face greater pressure to show that safety research and internal critique can coexist with rapid model development, rather than being treated as reputational liabilities.
Independent researchers and advocates gain a stronger reference point in debates over LLM governance, because concerns now span both former ethics staff and a foundational technical figure.
Third-order effects
If senior researchers increasingly voice concerns only after leaving major labs, AI governance may become more dependent on external scrutiny than on durable internal challenge mechanisms.
The episode points to a widening tension in AI industrialization: the institutions best positioned to build frontier systems must also retain credibility when their own researchers question the pace and safeguards of deployment.
The trend: AI’s rapid industrialization is making researcher independence and credible internal safety dissent central tests of major labs’ governance.
Interesting to read @demishassabis respond to @geoffreyhinton's resignation, saying he could have raised concerns + stayed within Google. Maybe - but linking such concerns to a potential DeSantis presidency might have been problematic - as he did to MIT. https://www.bloomberg.com…
@CadeMetz One of the most personally pressing issues is that this would have been a moment for Dr. Hinton to denormalize the firing of @timnitGebru (not to mention many that have recently followed). To say it was the wrong choice. Especially given his statements supporting ethica…
Sigh. I know some people who warned that Google, & AI broadly, would be dangerous to society. But we didn't collect $44M along the way & then shrug our shoulders about it. “The Godfather of AI' Quits Google and Warns of Danger Ahead - The New York Times https://www.nytimes.com/..…
I'm not arguing w the fact that AI poses risks. I AM ceaselessly annoyed by the pattern This is not new or novel. It was women - @timnitGebru @mmitchell_ai, me, et al - who rang the AI alarm years ago & were retaliated against, pushed out for doing so. https://www.nytimes.com/...
I asked @jaketapper to ask Hinton about his lack of support for @timnitGebru, @mmitchell_ai @mer__edith and other whistleblowers. His response is that their concerns were less existential than his and that it's easier to voice concerns if you leave first https://www.cnn.com/...
“People with much less power and more marginalized positions were taking real personal risks to name the issues with AI and of corporations controlling AI” @mer__edith on fire in this column from @parmy https://www.bloomberg.com/...
frustrations about the Dr Jeffrey Hinton interview in @nytimes - not least that the concerns he raises are the very same that @timnitGebru and team were silenced on and subsequently squeezed out of Google. Secondly the ‘too late for regulation’ mantra https://www.nytimes.com/...
The narrative of “I've earned my millions, now I think bigtech should consider ethics and privacy” is getting so repetitive, they should include the boilerplate text in contracts ... https://twitter.com/...
I'm so glad folks are questioning the redemption tours of people who step out of tech. Being an arsonist and getting lauded for warning of a fire you started is ridiculous. https://twitter.com/...