All news articlesNMC RESEARCH

Two Google DeepMind staff resign over military AI work

The departures follow a Pentagon deal allowing Google’s Gemini models into classified settings. The former staff question whether the company can oversee how its AI is used.

AI-generated editorial illustration based on the published article topic; not a documentary photograph.
Loading statistics…

Review information is loading.

Two Google DeepMind employees have resigned over the company’s military AI work, the BBC reports, bringing fresh attention to oversight of a deal that allows its technology to be used in classified Pentagon settings. [1]

Research engineer Sofia Liguori resigned in the week of 21 October; AI researcher Andreas Kirsch had left earlier that month. The BBC said it had also spoken to other workers who were leaving or considering leaving, but did not give a total for resignations. [1]

Speaking in a personal capacity, Liguori said Google’s growing push for AI systems in military operations made her “feel ashamed to work at Google”. She said the company’s work with the US and Israeli militaries had led her to lose faith in its stated ethical values. [1]

Kirsch told the BBC he believed Google’s oversight was insufficient for safely deploying AI in military and classified settings. In his view, the company appeared to have limited visibility into how the Pentagon used its technology and no right to control or veto that use. [1]

Google confirmed to Dutch broadcaster NOS in April that its Gemini models could be used in the Pentagon’s classified environment. The BBC reports that the agreement permits “any lawful government purpose”, but says the models “should not be used for domestic mass surveillance or autonomous weapons (including target selection) without appropriate human oversight and control”. [2] [1]

A Google spokesperson told the BBC that the company follows a consistent process for reviewing public-sector contracts against its use policies and AI principles. The spokesperson also pointed to channels for staff to raise concerns and policies against retaliation. Google’s published principles call for appropriate human oversight and alignment with international law and human rights. [1] [3]

The accounts do not establish which specific military operations, if any, have used Gemini, or how much oversight Google can exercise over its use in classified settings.

Sign in to give this article a thumbs up or down.

From the discussion

Private test discussion. Comments are readers’ views and are not yet automatically fact-checked. Editing is available for 60 seconds after posting.

Sign in with a confirmed reader account and choose a username to read comments and participate.

Sorting applies to top-level comments; replies remain oldest first. New comments and likes can change the order. Refresh for the current ranking.

Loading comments…