Several Google DeepMind employees have quit over the company’s military AI contracts, and others are reportedly on their way out. The BBC spoke to the workers in late October 2026. Their complaints boil down to one question: who decides how the Pentagon uses Google’s AI once the contract is signed?
Here’s who left, what the contract allows, and what Google says about it.
Why did Google DeepMind staff resign?
Google DeepMind staff resigned over the company’s military AI work, mainly a Pentagon contract signed earlier this year that lets Google’s AI be used in classified settings. The departing employees say they doubt there is enough human oversight, and that current AI models aren’t reliable enough to be used in warfare.
That’s the short version. The details are more interesting.
Who has resigned
Sofia Liguori is a London-based research engineer who spent three years at Google. She worked on AI for electrical-grid routing, according to The Intercept. She told the BBC that watching her company push AI deeper into military operations made her ashamed to work there. Her protest started earlier than her resignation: the day after Google announced its classified Pentagon deal in April, she took a personal sick-out, combining sick leave and holiday time. In her farewell letter to colleagues she told them to “stay human.”
Andreas Kirsch, an AI researcher, left earlier in October. He told the BBC he doesn’t think Google’s governance and oversight are strong enough for military and classified work. He also argues that today’s AI models are not reliable enough for warfare.
René Mayrhofer is a separate case. He’s a director for Android platform security, not a DeepMind researcher, and LBC reports that his farewell letter called his decision unavoidable. It shows the unease isn’t limited to one team.
How many people have actually left? The BBC says “several” and doesn’t give a number. Some headlines call it an exodus. The reporting doesn’t support a figure either way.
What the Pentagon contract allows
According to the BBC, Google’s agreement lets the US military use its AI for classified work for “any lawful government purpose.” The agreement also says the technology shouldn’t be used for domestic mass surveillance, or for autonomous weapons (including target selection), without appropriate human oversight.
Kirsch’s objection is about control. In his view Google seems to have no right to veto how the Pentagon uses its AI, and little visibility into how it’s deployed.
The timing matters too. US Defense Secretary Pete Hegseth said in January that the US would become an “AI-first warfighting force.” Emil Michael, an under secretary at the Department of Defense, has said speed defines victory in the AI era. The Pentagon’s argument is that adopting AI quickly is essential.
One example of military AI already in use is Maven Smart System, the Pentagon’s main software platform for organising intelligence and data. Palantir builds it and has plugged in AI from several leading AI companies.
Which other companies signed Pentagon AI deals?
Google isn’t alone. The BBC reports that OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia and the start-up Reflection signed similar contracts, covering military uses such as transport, planning and battlefield operations.
That’s relevant if you’re wondering whether Google is an outlier. On the contracts themselves, it isn’t. The pushback inside DeepMind is the part that stands out. If you follow how AI companies write their rules, our piece on the Anthropic usage policy update shows how much those policies are shifting this year.
The Project Nimbus background
This isn’t the first time Google staff have objected to military-adjacent work. In 2024, employees criticised Project Nimbus, a joint contract with Amazon worth about $1.2bn that supplies Israel’s government and military with AI and cloud infrastructure. Google said at the time that the work wasn’t aimed at highly sensitive, classified or military workloads tied to weapons or intelligence services.
Liguori’s resignation covers both the US and Israeli contracts, per The Intercept.
Can Google employees speak out?
People inside Google told the BBC that staff are afraid to speak publicly about AI safety and warfare, because their employment contracts prohibit criticising the company. Kirsch said the same on X.
One current DeepMind employee, who stayed anonymous for fear of retaliation, told the BBC that many motivated people may leave, but that there’s also a steady supply of new hires who’ll fill the roles. The Intercept adds that DeepMind employees voted to unionize in May, and that some colleagues are organising against the contracts.
What Google says
A Google spokesperson told the BBC the company works with public sector organisations worldwide and follows a consistent process for reviewing contracts, including checks against its acceptable use policies and AI principles. Google also says it has dedicated channels for employees to raise concerns, backed by strict anti-retaliation rules.
Those AI principles include human oversight and respect for international law and human rights. The resigning employees say the oversight in practice isn’t strong enough, which is exactly where the two sides disagree.
Google’s AI ambitions are only growing. We’ve covered why Demis Hassabis says Gemini 3.0 Pro left ChatGPT behind and the next model Google is testing, Gemini 4 Carbon. More capable models will keep raising the same oversight question.
FAQ
How many Google DeepMind employees have resigned?
The BBC reports that several workers have resigned or are in the process of leaving. It doesn’t give a total.
Does Google’s Pentagon contract allow autonomous weapons?
Per the BBC, the agreement says Google’s AI shouldn’t be used for autonomous weapons, including target selection, without appropriate human oversight. Critics point out that the contract also permits “any lawful government purpose,” and that Google has limited visibility into deployment.
What has Google said about the resignations?
Google says it reviews government contracts against its acceptable use policies and AI principles, and that employees have channels to raise concerns without retaliation.
Is this the first employee backlash at Google over military work?
No. Staff also criticised the 2024 Project Nimbus contract with Israel’s government and military.
