Google DeepMind Researchers Quit Over Military AI Contracts

Some of the people building the world’s most advanced AI are walking away over how it might be used in war. Several researchers at Google DeepMind, the company’s AI lab, have resigned or are preparing to leave in protest over Google’s contracts with the US military and Israel. Their concerns centre on whether Google can control what happens to its technology once it reaches military hands. The Google DeepMind resignations have reignited a long-running debate about military AI ethics and the role of big tech in defence. Here is what happened and what both sides say.

Who Has Resigned

The departures were reported in a BBC investigation published on 21 October 2026. Research engineer Sofia Liguori, who joined DeepMind in London in 2023 and worked on AI to help route electricity through power grids more efficiently, resigned this week after three years. AI researcher Andreas Kirsch left earlier in October, as the BBC’s reporting summarised explains.

Kirsch told the BBC he believes Google’s governance and oversight are not sufficient or meaningful enough to deploy AI safely and responsibly in military and classified settings. Liguori said her internal concerns went nowhere before she decided to leave. Other Google staff, including a senior Android security director, have also announced departures over similar concerns.

The Contracts at the Centre

Two arrangements are drawing criticism:

• A Pentagon agreement signed earlier this year that allows Google’s AI models to be used across classified military environments for any lawful government purpose. The Pentagon named Google among eight companies supplying AI for classified networks in May.
• Project Nimbus, a US$1.2 billion cloud and AI contract shared with Amazon that serves Israel’s government and military, which has faced staff protests since 2024.

The Pentagon contract reportedly bars autonomous weapons and mass surveillance without human oversight. Critics argue those limits are hard to enforce once technology is deployed in classified settings, where outside visibility is limited.

Google’s Response

Google rejects the idea that it has abandoned its principles. A spokesperson told the BBC that the company follows a consistent process for public sector contracts, complies with its AI Principles and gives employees channels to raise ethical concerns without retaliation, as Ynet News reported.

Supporters of defence partnerships argue that democratic governments will use AI regardless, and that responsible companies should help ensure it is used with strong safeguards rather than leave the field to others.

Why the Debate Matters

The disagreement goes to the heart of how AI should be governed. Key questions include:

• Control after delivery. Once a model is handed to a military, how much say does the company keep over its use?
• Reliability. AI systems can make mistakes. In military settings, errors can have life-or-death consequences.
• Speaking up. Kirsch has said staff are afraid to criticise the company publicly because of employment contracts, which raises concerns about how dissent is handled.

DeepMind’s history adds weight. When Google acquired the lab in 2014, it reportedly agreed to conditions that included limits on military use. Critics, including Kirsch, argue those commitments have not survived.

For more on AI policy and safety, see our reports on the White House Super Intelligence task force and Google’s Gemini 4 Argon. Our AI ethics explained guide covers the wider debate.

A Test of Principles

The DeepMind resignations show that the people building AI are increasingly willing to act on their concerns about how it is used. Google insists its safeguards are strong, while critics say oversight ends at the military’s door. However this dispute unfolds, expect pressure for clearer, enforceable rules on military AI to keep growing.

Frequently Asked Questions

Who resigned from Google DeepMind?

The BBC named research engineer Sofia Liguori and AI researcher Andreas Kirsch, among others who have left or plan to leave.

Why did they resign?

They say Google’s oversight is not strong enough to ensure its AI is used safely and responsibly in military and classified settings.

What does Google say?

Google says it follows its AI Principles, reviews public sector contracts consistently and lets staff raise concerns without retaliation.

Does the Pentagon deal allow autonomous weapons?

Reports say the contract bars autonomous weapons and mass surveillance without human oversight, though critics question how that is enforced.