Several Google DeepMind employees have resigned over Google's growing military AI work, according to a BBC investigation that has since been widely reported. The departures put a spotlight on a question the whole industry faces: what control does an AI company keep over its technology once it is deployed in classified government settings?
What happened
Sofia Liguori, a London-based research engineer who spent three years at DeepMind, resigned and spoke publicly. She told the BBC that realising her company was pushing for AI in military operations made her feel ashamed to work at Google. AI researcher Andreas Kirsch also resigned earlier this month. According to reports, other employees interviewed anonymously voiced similar concerns, and some are said to be preparing to leave.
The resignations follow Reuters reporting that Google signed a classified agreement with the US Department of Defense earlier this year. The deal reportedly allows its AI models to be used for "any lawful government purpose," including classified military work. In June, René Mayrhofer, a director for Android platform security, also resigned, citing Google's work with the defence department. In May, DeepMind workers had called for union recognition over the use of their research by the US and Israeli militaries.
The employees' concerns
According to the BBC's reporting, the staff who spoke out worry about three things:
- Whether Google can see how its models are used once they run inside classified environments.
- Whether human oversight can stay meaningful when AI speeds up decisions such as battlefield targeting.
- Whether the technology is reliable enough for warfare without more testing and stronger guardrails.
Kirsch said Google appears to have limited ability to control or veto how the Pentagon uses its AI and limited visibility into deployment. One outlet also reports that the contract requires Google to adjust AI safety settings at the government's request, a detail that comes from secondary coverage and that Google has not been quoted confirming in the material reviewed.
Google's response
A Google spokesperson said the company works with public-sector organisations worldwide and follows a consistent process for reviewing and entering contracts, including compliance with its acceptable use policies and AI principles. It added that employees have dedicated channels to raise concerns and that strict policies against retaliation are in place.
The wider context
Google is not alone. Coverage notes that OpenAI, Microsoft, Amazon and other firms have pursued or signed similar agreements to deploy AI across classified military environments. US defence leadership said in January that the country intends to become an "AI-first" warfighting force. Supporters of such deals argue that democratic governments need advanced AI for national security and that lawful use is already governed by existing rules. Critics argue the safeguards are too thin when systems operate in secret.
Why it matters
The resignations are small in number but significant for the industry's biggest research lab, because they show talent, not only outside critics, pushing back on how frontier AI is used. How companies balance government contracts with their own published AI principles, and what they tell employees and the public about it, is likely to remain a live issue.
Sources: BBC investigation as reported by Dig.Watch, LBC and IBTimes; Reuters reporting as summarised by Gadget Review.
