Google Military AI Contract Sparks DeepMind Employee Resignations

Google military AI contracts trigger wave of resignations

Image source, Reuters/supplied

Image caption, Sofia Liguori resigned from Google, saying the company's push to use more AI technology for military purposes made her feel "ashamed" to work there. Article Information

Author, Christopher Giles

Role, BBC Eye Investigations

Published 9 minutes ago

Reading time: 4 minutes

Google is facing a backlash from some of its employees after deploying its artificial intelligence (AI) technology for military use.

BBC has spoken to several tech workers who have either recently resigned from Google's AI company Google DeepMind or are preparing to leave, citing concerns over the tech giant's military contracts.

London-based research engineer Sofia Liguori is one of them. She resigned from Google DeepMind this week and spoke exclusively to the BBC about her concerns, despite her contract restricting her from doing so.

Against the backdrop of US-China AI militarization, BBC learns the Pentagon has formally discontinued Anthropic tools October 6, 2026

BBC steps into the frozen lair of Google's quantum computer — a rival to China's "Zuchongzhi 3.0"? January 11, 2026

The era of robots fighting each other is here? The battlefield in Ukraine foreshadows a disturbing future of war May 7, 2026

End of content

A Google spokesperson told the BBC: "We work with public sector agencies around the world and follow a consistent process to review and sign these contracts, including compliance with our Acceptable Use Policy and AI Principles."

"We also have dedicated channels for employees to safely raise any concerns, covering everything from business conduct to misconduct, and we have strict workplace policies in place to prevent any form of retaliation."

Google is one of several tech companies that signed contracts with the US military earlier this year to deploy AI technology in classified military environments. US Defense Secretary Pete Hegseth has made integrating AI into the armed forces a top priority, and said the Trump administration will become an "AI-first fighting force."

Image source, Getty Images

Image caption, US Defense Secretary Pete Hegseth said in January that the US "will become an 'AI-first' fighting force."

Looking at UK statistics, why are people who have never smoked also becoming addicted to e-cigarettes?

North Korea releases photos of Kim Jong Un: why do rumors of his death always resurface?

After three years at the company, Liguori said she decided to resign because of Google's deepening cooperation with the US and Israeli militaries. "I don't agree with this direction, and it has also cost me my confidence that the company can stay true to the ethical values it claims to uphold," she said.

According to Google's AI Principles, these values include incorporating "human oversight" in the implementation of AI and making the technology consistent with "international law and human rights." But current and former employees interviewed by the BBC, as well as relevant experts, said they are concerned whether the level of human oversight at critical moments is sufficient.

These tech workers said they are especially worried about the AI military agreements signed this year, because these agreements significantly expanded the scope of technology the Pentagon can use.

AI researcher Andreas Kirsch also resigned from Google DeepMind earlier this month over related concerns.

Kirsch told the BBC he believes "Google's governance and oversight mechanisms are inadequate, or lack sufficient meaning, to safely and responsibly deploy AI in military and classified environments," and are also insufficient to "safely develop artificial superintelligence in the future."

He added: "Google appears to have no power to control or veto how the Pentagon uses its AI, and it also appears unable to fully grasp how it is deployed."

Image source, Supplied

Image caption, former Google employee Kirsch believes existing AI models are not yet "robust" enough for military use.

Over the past few years, the US military has deployed a number of AI-driven technologies to enhance its combat capabilities.

The most famous example is the Maven Smart System, the Pentagon's main software platform for organizing intelligence and data.

The system was developed by US defense company Palantir and uses AI to help personnel select and identify strike targets. Palantir has integrated technology from several leading AI companies into the Maven system, which has significantly accelerated some processes.

Lt Gen (Ret) Jack Shanahan, a former head of Project Maven, said the current "philosophy" appears to be: "deploy (AI technology) as fast as possible and use it as widely as possible, even if that may come at the cost of the rigour and discipline required for testing and evaluation."

However, the Pentagon has consistently insisted that adopting AI technology is crucial to winning wars. US Deputy Secretary of Defense Emil Michael said in a statement earlier this year: "Speed determines victory in the AI era."

The contract Google and the Pentagon signed earlier this year allows its AI technology to be used for classified military projects. The agreement states that Google's AI models may be used for "any lawful government purpose" and that they "should not be used for domestic mass surveillance or autonomous weapons (including target selection) without appropriate human oversight and control."

Image source, Justin Sullivan/Getty Images

Image caption, Google's AI Principles include incorporating "human oversight" when implementing AI and aligning it with "international law and human rights."

OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia and the startup Reflection have also signed similar agreements allowing AI technology to be deployed in a range of military scenarios, including transport, planning and battlefield operations.

As the technology develops rapidly, concerns about the use of AI on the battlefield are also growing.

Sarah Shoker, a former head of OpenAI's geopolitical team, said: "AI operates far faster than humans, so it compresses the time operators have to make decisions, which is especially dangerous in high-pressure environments."

"When hundreds of targets are selected by AI and the time available for decisions is short, it is almost impossible for human operators to genuinely review AI outputs for errors."

Google employees told the BBC that, because their employment contracts prohibit them from publicly criticising the company, staff are afraid to speak out about AI safety and its use in warfare.

Sophia Liguori told the BBC: "In principle, I would never be allowed to criticise Google for my whole life." But she decided to do so, "because I think it is important for the outside world to know these things, and I also think that stating facts is not the same as criticising."

Liguori added that at Google, "people are afraid of retaliation, afraid of it affecting their careers. You feel like you are being watched and cannot speak out."

Audio caption, drones dominate today's battlefield, and AI acting as a fighter jet "pilot" is also becoming a future trend.

A current Google DeepMind employee, who declined to be named for fear of retaliation, told the BBC: "Many motivated employees who joined because they identified with the (company's) original mission may leave, but at the same time there is an endless stream of new hires who can fill those positions. This may be a generational change."

Another Google DeepMind employee, who also wished to remain anonymous, told the BBC: "Many people who feel strongly about this matter (how Google DeepMind has changed its ethical stance) have already left the company."

The employee said: "Some are still fighting, while others are thinking: 'This is a big company. Do you really expect to be able to change anything?'"

Google's agreement with Israel has also sparked controversy. In 2024, the company was criticised by employees for signing a $1.2bn (£960m) partnership contract with Amazon called Project Nimbus. The project provides AI technology and cloud computing infrastructure to the Israeli government and military. At the time, Google said Project Nimbus was "not intended for highly sensitive, classified or military workloads related to weapons or intelligence services."

In addition, Israel's AI capabilities in the war have also become a focus of discussion, because the recent documentary "Naza" ("Kill Code: NAZA") explored the role of AI-driven target selection systems in the bombing of the Gaza Strip and pointed out that they caused a large number of civilian casualties. An Israeli government minister said the claims in the film were false and misleading.

Former Google employee Andreas Kirsch believes that military AI technology is not yet ready for the battlefield. He said: "Existing AI models are still not robust enough, and I think they have not been trained for this kind of task either."

Additional reporting by Rosa Rahimi

This article was originally written in English. We used AI to assist with the translation, and it was reviewed by BBC journalists before publication. Learn more about how we use AI.