Several Google DeepMind researchers have resigned in protest over the company's military artificial intelligence contracts with the United States and Israeli militaries, the BBC reported on Wednesday, reopening a bitter internal fight over AI ethics at one of the world's most influential AI labs.
 |
| File photo: Google DeepMind's headquarters at 6 Pancras Square, London. (Wikimedia Commons) |
The resignations
London-based research engineer Sofia Liguori quit this week after three years at the company, and spoke to the BBC even though she says her contract restricts her from criticizing Google publicly. "Realising my company is pushing more and more for AI systems in military operations makes me feel ashamed to work at Google," Liguori told the BBC in a personal capacity. "I don't agree with this direction and this has led me to lose faith in my company being true to its stated ethical values."
AI researcher Andreas Kirsch also resigned from Google DeepMind earlier this month, telling the BBC that Google's governance and oversight are "not sufficient or meaningful enough to deploy AI safely and responsibly in military and classified settings", or "to develop artificial superintelligence safely in the future". His sharpest point was about control: "Google seems to have no right to control or veto how the Pentagon uses its AI and also seems to have limited visibility into how it's being deployed."
Two current DeepMind employees, speaking anonymously, told the BBC that "a lot of people have left the company that feel strongly about this". And the departures did not start with Liguori: in September, engineer Robert O'Callahan quit because his team was making AI run faster, the first reported principled exit of the autumn.
 |
| File photo: The Pentagon, which signed classified AI contracts with Google and other tech firms earlier this year. (Wikimedia Commons) |
What Google signed
At the center of the revolt is a Pentagon contract signed earlier this year. In May, the Defense Department named Google among eight companies agreeing to provide AI capabilities for classified military networks, according to the BBC. Google's agreement allows its models to be used for "any lawful government purpose", while saying they "should not be used for domestic mass surveillance or autonomous weapons (including target selection) without appropriate human oversight and control". OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia and the start-up Reflection signed similar contracts.
The push comes from the very top of the Pentagon. Defense Secretary Pete Hegseth said in January that the United States will become an "AI-first warfighting force", and the Defense Department has since set up a command for autonomous warfare while shopping for AI agents capable of hacking on their own. The stakes were underscored when Anthropic's Claude was dropped from Palantir's Maven Smart System — the Pentagon's main targeting platform — after Anthropic refused to lift its limits on military use.
Google's own AI Principles once promised more. The 2018 version listed categories of AI "we will not pursue", including weapons and surveillance that violate international norms. Google dropped that list when it rewrote the principles in February 2025, in a blog post co-written by DeepMind chief Demis Hassabis. The current version promises "appropriate human oversight" and alignment with "widely accepted principles of international law and human rights" — language the resigning engineers say is no longer enough.
A farewell letter, and fear inside the lab
Liguori did not leave quietly. She has since published a farewell letter, "Stay Human", dated October 20 — her last day at Google DeepMind. Writing in a personal capacity, she says she had been on leave since Google signed its classified contract with the Pentagon in April, after a year of campaigning against the company's work with Israel under the Project Nimbus cloud contract, with defense contractor Lockheed Martin, and with US immigration and border agencies. Liguori joined DeepMind in 2023 and had helped organize a demonstration outside its London office with the activist group No Tech for Apartheid.
Her main charge is about how such deals are written. She says the contracts allow AI to be built into military and surveillance systems "while imposing no restrictions on permissible uses and leaving the company no monitoring power" — which she calls "a very convenient contract design for executives to escape accountability". She also describes a climate of fear inside the lab: staff who raised military contracts with their teams, as she did, got "called into HR meetings on the pretext of having breached some internal policy", which she calls intimidation. "In principle I'm not allowed to criticise Google for the duration of my life," Liguori told the BBC, adding that "talking about facts is not criticising". Inside the company, she said, "people are scared of retaliation and putting their career at risk."
She does not urge others to follow her out. "I don't think resigning is an effective method, certainly not as an individual act," she writes, asking colleagues who stay to join the GDM Union, through which UK-based DeepMind staff are organizing with the UTAW and Unite unions. Its demands, she says, include "enforceable red lines on military/surveillance applications", more transparency around military contracts, the right to abstain from work on ethical grounds, and a stronger sustainability commitment.
Google defended its process. A spokesperson told the BBC: "We work with public sector organisations worldwide and follow a consistent process for reviewing and entering these contracts, including compliance with our acceptable use policies and AI principles." The company added that it has "dedicated channels for employees to safely raise any concerns" and "strict workplace policies in place to prevent retaliation of any kind". It has not commented on Liguori's farewell letter.
Echoes of Project Maven
Google has been here before — and last time, the dissenters won. In 2018, thousands of employees protested the company's work on the original Project Maven, a Pentagon effort to analyze drone footage with AI. Some resigned, Google let the contract lapse, and it published those first, stricter AI principles. In 2024, its $1.2 billion Project Nimbus cloud deal with the Israeli government, shared with Amazon, drew fresh staff protests.
But this time the company is staying in, and the dissenters are walking out. Retired lieutenant general Jack Shanahan, who led the original Project Maven, told the BBC that the philosophy now seems to be to "go as fast as possible to deploy [AI technology] as widely as possible, potentially at the expense of the rigour and discipline of test and evaluation". If the people most worried about how DeepMind's models are used keep leaving, fewer voices inside the lab will be left to push back — just as those models move into classified war planning.
What This Means
The resignations expose a structural collision between the AI industry's founding ethics culture and the Pentagon's drive to weave AI into warfare. The contract at issue does not merely buy Google's models for the military — its sweeping "any lawful government purpose" language is precisely what terrifies the engineers who built them, because, as Kirsch argues, it hands the Pentagon wide latitude while leaving Google little visibility into how the technology is deployed and no clear power to block specific uses.
The contrast with 2018 is stark. Then, employee pressure pulled Google out of military AI; now, the company is holding the line while the objectors resign — suggesting the balance of power between Big Tech's workers and its executives has shifted. It also puts Google in a visible split with Anthropic, which refused to loosen its military-use limits and lost its place on the Pentagon's main targeting platform for it. For the broader industry, the episode is a warning sign: as AI companies chase some of the largest defense contracts in history, they risk losing the researchers whose trust and talent built their labs — and whose warnings about safety were meant to be the guardrails.
 |
| File photo: Defense Secretary Pete Hegseth, who has made building an "AI-first warfighting force" a top priority. (Wikimedia Commons) |
What Happens Next
The immediate question is whether Liguori's exit is the end of the story or the beginning of a larger one. Two current DeepMind employees told the BBC that more staff have left quietly, and her call for colleagues to join the GDM Union could turn individual resignations into organized pressure with concrete demands — enforceable red lines on military and surveillance applications, contract transparency, and the right to refuse work on ethical grounds.
Google, for its part, has shown no sign of retreating: its Pentagon contract remains in force, and Hegseth's "AI-first warfighting force" doctrine is accelerating deployments. What bears watching is whether other engineers follow the same path, whether the union push gains traction among DeepMind's UK workforce, and whether the debate spills into Washington — where lawmakers have shown growing interest in how commercial AI models are used in classified settings. With midterm elections less than two weeks away and defense AI now a mainstream political issue, the fight that began inside DeepMind's London office may not stay there.
Sources: BBC News interviews with Sofia Liguori and Andreas Kirsch and Google's statement (Oct 21, 2026); Sofia Liguori, "Stay Human" farewell letter (Oct 20, 2026); Google AI Principles; The Neuron AI digest (Oct 22, 2026).
Post a Comment