Google likes to portray itself in the AI field as the reliable veteran: steady, experienced, and grounded in long-term thinking. Even amid its biggest AI organizational overhaul announced on Wednesday, company leaders stuck to a consistent narrative about preparing for stronger results ahead.

Yet the string of leadership moves points to deeper strains inside Google’s AI teams. These include its uncertain standing in the competitive AI landscape, Demis Hassabis’ preference at DeepMind for extended research rather than rapid product releases, and growing internal pushback from senior figures such as Jeff Dean.

One current Google staff member, speaking anonymously over concerns about reprisals, described Dean’s exit as the end of the company’s earlier idealistic phase. “Losing Jeff adds to the talent exodus at Google,” the employee noted. “Average performance is the probable outcome.”

The announcements surprised the AI community, where researcher movement between organizations has become routine. Google revealed that Hassabis would step back from daily DeepMind duties to concentrate on AGI and core research, with CTO Koray Kavukcuoglu taking over leadership. Meanwhile, longtime employee Jeff Dean and three senior AI researchers are departing to start a new AI venture.

“Losing Jeff adds to the talent exodus at Google. Average performance is the probable outcome.”

Google is seen as the strongest performer among legacy tech companies in AI, outpacing Meta, Amazon, and Apple. Still, it trails OpenAI and Anthropic in speed and has not yet overtaken them on cutting-edge models. Both Hassabis and CEO Sundar Pichai highlighted this challenge in their remarks.

“We remain dedicated to staying at the leading edge and are sharply focused on areas needing improvement,” Pichai stated, noting the need to “keep advancing quickly and with clear intent.” Hassabis described the moment as a “new phase” where the company possesses “everything required to take the lead going forward.”

Industry observers interpreted the moves as Hassabis yielding ground to emphasize quicker, product-oriented progress at Google DeepMind.

“Internal rivalry around AI products has grown more intense than at any point in the past ten years,” a former Google employee with direct knowledge of operations told The Verge. The source noted rising pressure to recruit staff capable of accelerating product launches amid tougher competition.

“Internal rivalry around AI products has grown more intense than at any point in the past ten years.”

A separate industry source, requesting anonymity due to ongoing lab collaborations, remarked, “This reads as a clear reduction in Demis’ sway.”

Hassabis has often emphasized scientific breakthroughs from AI over immediate commercial applications. Pichai indicated Hassabis will now focus on “guiding the direction of AGI” in his new roles while continuing to lead Isomorphic Labs.

AI leadership represents only part of Google’s challenges. Sources repeatedly raised questions about how much the changes stem from mounting ethical disputes inside the company.

Google’s contract with the Department of Defense may have contributed to both Hassabis’ adjustment and Dean’s exit, according to Tyler Johnston of The Midas Project. The agreement could enable Google AI use in widespread surveillance or autonomous lethal systems.

“This has already triggered staff exits and union activity at DeepMind,” Johnston said. “Signs suggest Dean and Hassabis held personal objections. DeepMind’s original acquisition included a pledge against military or intelligence applications.”

“DeepMind’s original acquisition included a pledge against military or intelligence applications.”

Blogger Zvi Mowshowitz suggested Dean’s departure might connect to the Defense Department contract decisions.

Dean plans to launch Discovery Loop, an AI startup focused on automating scientific and engineering experiments in drug discovery, hardware, clean energy, and materials. Cofounders include Sanjay Ghemawat, Quoc Le, and Oriol Vinyals.

“Losing Jeff represents a setback for ethics and integrity at Google,” the current employee said. “He stood out as one of the few senior voices willing to address concerns like surveillance and autonomous weapons.”

Dean had been a prominent critic of certain military AI policies and supported legal challenges against restrictions on companies maintaining ethical boundaries. He also publicly condemned specific government actions on immigration enforcement.

Dean had previously voiced opposition to the same ethical boundaries as Anthropic.

Following Anthropic’s conflict with the Pentagon, Google agreed to permit its AI for any lawful Pentagon purpose, filling the space left by stricter policies elsewhere.

Dean had earlier criticized mass surveillance and lethal autonomous weapons in public statements and signed related pledges alongside Hassabis.

Former Google researcher Alex Turner argued in a recent piece that Dean’s exit would harm the company and noted the tension created by his public positions.

Are you a current or former Google employee? Contact securely via Signal at haydenfield.11

In the amicus filing Dean joined, the signers stressed their expertise in large-scale AI systems and the risks when deployment outruns legal and ethical safeguards.

While Dean’s departure may quiet one prominent voice, broader discontent persists. Employees have voiced growing frustration with company political directions this year.

“Leadership has shifted its stance dramatically, and now another key figure willing to raise moral concerns has left,” the current employee said. “It raises questions about future compromises.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

Hayden Field