Google's AI shakeup signals deeper rifts over Pentagon deal and ethics

Google announced a major reshuffling of its AI leadership on Wednesday that, despite the company's carefully unified public messaging, masks deep internal divisions over ethics and the pace of AI development. Demis Hassabis, long the head of DeepMind, is stepping away from day-to-day operations to focus on AGI and research, moving into roles as DeepMind chair and Alphabet's chief scientist while maintaining his position at Isomorphic Labs, the drug-discovery spinoff. Koray Kavukcuoglu, DeepMind's current CTO, will assume leadership of the division.

The more seismic announcement involves Jeff Dean, a 27-year Google veteran and the company's 30th employee, who is leaving to launch Discovery Loop, an AI startup focused on automating thousands of science and engineering experiments in drug discovery, hardware design, clean energy, and material design. He is joined as cofounders by three additional top Google AI researchers: Sanjay Ghemawat, Quoc Le, and Oriol Vinyals. The departures come amid broader friction within Google's AI organization on multiple fronts. Google CEO Sundar Pichai stated the company is "committed to being at the frontier" and must "continue to move fast," while Hassabis wrote that the company is entering a "next chapter." Yet industry sources and current employees frame these changes as evidence that Hassabis has lost influence in favor of a faster-moving, product-focused leadership eager to close gaps with OpenAI and Anthropic, which are viewed as more nimble than Google despite Google's own acknowledged edge over Meta, Amazon, and Apple.

Underlying the reshuffling, however, is a deeper ethical fault line that appears to have pushed Dean out. In April, Google signed a Pentagon contract permitting AI use for "any lawful use"—language that potentially allows deployment in mass surveillance and lethal autonomous weapons. Dean had been one of the company's most visible critics of such uses. In March, he co-signed an amicus brief supporting Anthropic's lawsuit against the Department of Defense after Anthropic was blacklisted from Pentagon contracts for refusing to drop its red lines: prohibiting domestic mass surveillance and lethal autonomous weapons without human control. In February, Dean wrote on X that "Mass surveillance violates the Fourth Amendment and has a chilling effect on freedom of expression," and in 2018, both Dean and Hassabis signed a public pledge stating "the decision to take a human life should never be delegated to a machine." The irony is stark: DeepMind was itself acquired by Google under an agreement not to sell the technology for military or intelligence purposes—a commitment Google appears to have abandoned.

Google researcher Alex Turner, who quit over the Pentagon deal, wrote in an op-ed that Dean's departure "would be a disaster for the company" and that "by signing [the amicus brief], he broke from the company line, which probably made things awkward within the C-suite." Tyler Johnston, founder of The Midas Project, an AI industry watchdog nonprofit, told The Verge that the Pentagon contract "has already caused staff departures and a unionization effort at DeepMind" and that "there are many signs suggesting Dean and Hassabis may have personally felt opposed." A current Google employee, who requested anonymity due to fear of retaliation, described Dean's departure as exemplifying "the vestiges of Google's idealism" fully dying out: "Losing Jeff continues the brain drain at Google. Mediocrity is the likely result." The same employee noted the broader erosion of ethical guardrails: "Sergey [Brin] did a 180 politically, Sundar stood behind Trump at his inauguration. Now one of the only remaining figures at the company willing to speak out on moral issues is gone. Who knows what we will sell out next?"