For brands selling artificial intelligence, trust is the product. That is why the resignations at Google DeepMind matter beyond the technology press: when the researchers who build a system publicly quit over where it is used, every customer pitch that leans on responsible AI gets harder to make.
Several DeepMind researchers have resigned over Google’s military AI contracts, according to a BBC investigation. Sofia Liguori, a London based research engineer, resigned after three years, saying the push into military operations made her feel ashamed to work at Google. Andreas Kirsch resigned earlier in October, arguing that governance and oversight were not sufficient for classified military settings.
The trust problem for AI sellers
Enterprise AI is sold on safety, oversight and control. Google’s Pentagon agreement allows its models to be used for any lawful government purpose in classified operations, with stated restrictions on domestic mass surveillance and autonomous weapons without appropriate human oversight. Kirsch’s complaint cuts directly at that pitch: he said Google seemed to have no right to control or veto how the Pentagon uses its AI, and only limited visibility into its deployment.
For marketers and buyers, the gap between a stated principle and a visible resignation is the story. A company can publish AI principles, run responsible AI pages, and brief customers on safeguards, but a named researcher quitting in public reframes all of it as marketing copy rather than governance. Two current staff, speaking anonymously, told the BBC that many who felt strongly had already left, which suggests the pipeline of internal dissent is emptying into public view.
What buyers should ask
Google said it reviews public sector contracts against its acceptable use policies and AI principles, and that employees have channels to raise concerns without retaliation. That is the standard corporate answer, and it will not settle the question for buyers who now have to explain the headlines to their own boards.
As the Financial Express reported, the resignations follow months of internal opposition, including a petition signed by around 600 employees in April urging leadership to stay away from classified military work.
The practical lesson for any AI brand is that talent is a trust signal. When researchers stay, principles look credible. When they leave and say why, principles look like positioning. The firms that win enterprise AI deals in the next year will be the ones whose own builders are willing to stay and defend the work in public.
