Several researchers at Google DeepMind have resigned over the company’s military artificial intelligence contracts, according to a BBC investigation reported this week, in the clearest sign yet that the ethics debate inside big technology firms has moved from petitions to departures.
Sofia Liguori, a London based research engineer, resigned after three years at the company. She told the BBC that realising her company was pushing more and more for AI systems in military operations made her feel ashamed to work at Google. Andreas Kirsch, an AI researcher, resigned earlier in October, arguing that Google’s governance and oversight were not sufficient for deploying AI safely in military and classified settings.
What the contracts allow
Google signed an agreement with the Pentagon earlier this year permitting its AI models to be deployed in classified government operations. The contract allows use for any lawful government purpose, though Google says its models should not support domestic mass surveillance or autonomous weapons, including target selection, without appropriate human oversight and control.
Kirsch said Google seemed to have no right to control or veto how the Pentagon uses its AI, and only limited visibility into how it was being deployed. Two current DeepMind staff, speaking anonymously, told the BBC that many people who felt strongly about the company’s changing ethics position had already left. In April, around 600 employees signed a petition urging leadership to stay away from classified military work.
Google’s answer
Google said it reviews public sector contracts for compliance with its acceptable use policies and AI principles, and that employees have dedicated channels to raise ethical concerns. The company also said its workplace policies prohibit retaliation against those who speak up, and that its military agreements include restrictions on mass surveillance and autonomous weapons.
As the Financial Express reported, the resignations follow months of internal opposition, and researchers who left argue the restrictions may not work as effectively in practice as Google claims once technology enters classified environments.
The departures raise a practical question for every firm selling frontier AI to governments: who watches the watcher once the system leaves the lab? Resignations do not stop contracts, but they do signal to customers and recruits alike that the people building these systems are no longer willing to stay quiet about where they end up.
