Google staff resign over DeepMind's military AI contracts

The BBC reported on 21 October that several people have resigned from, or are leaving, Google DeepMind over the company’s AI contracts with the military. It named two: Sofia Liguori, a London-based research engineer who left that week after three years, and Andreas Kirsch, an AI researcher who resigned earlier in the month. The account rests on the BBC’s interviews, including with two current DeepMind staff who were not named.

Liguori told the BBC she was “ashamed” that her employer was pushing AI into military operations and cited Google’s growing work with the US and Israeli militaries. Kirsch said he did not think Google’s governance and oversight were sufficient for deploying AI in military and classified settings, and that Google appeared to have no right to veto how the Pentagon uses its AI and limited visibility into deployment. The BBC said Google was one of several companies that signed contracts earlier this year to deploy AI on classified military networks, naming OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia and Reflection among those with similar agreements. It reported that Google’s contract allows use for “any lawful government purpose” while saying the AI should not be used for domestic mass surveillance or autonomous weapons, including target selection, without appropriate human oversight and control.

The second claim in the report concerned speech. Liguori said she spoke despite her contract, adding: “In principle I’m not allowed to criticise Google for the duration of my life.” Unnamed current employees told the BBC that staff are afraid to speak out on safety and warfare because their contracts forbid public criticism of the company, and Liguori said people fear retaliation. The BBC did not publish the contract text, and the claim about lifelong restrictions rests on Liguori’s reading and employees’ accounts; Google’s statement, as quoted, did not address it directly.

A Google spokesperson said the company works with public-sector organisations worldwide and follows “a consistent process” for reviewing contracts, including compliance with its acceptable use policies and AI principles. The spokesperson added that it has dedicated channels for raising concerns and strict policies against retaliation. The Pentagon’s position, as the BBC reported it, is that adopting AI is essential to winning wars; an under secretary of defence was quoted saying “Speed defines victory in the AI era”. Critics quoted, including a former director of Project Maven, questioned whether test and evaluation is keeping pace.

The resignations follow that of DeepMind safety researcher Alex Turner, who left over Google’s Pentagon agreement and wrote about it in July. OpenAI signed its own classified-network agreement in February. Because this is a single outlet’s account with mostly named-individual and anonymous testimony, the scale of departures is not established; the BBC gave no count.