Google DeepMind researchers resign over Pentagon AI contracts and ethical concerns

Why researchers resigned

Three Google DeepMind researchers—Sofia Liguori, Andreas Kirsch, and Alex Turner—have resigned in recent weeks, citing ethical concerns over the company’s AI contracts with the US Department of Defense (Pentagon). The agreements permit the Pentagon to deploy Google’s AI models in classified military environments, where the company’s oversight is limited and its ability to veto specific uses is unclear.

The departing researchers argue that once the models are integrated into air-gapped military networks—systems isolated from unsecured networks—the existing safeguards and human oversight become inadequate. They emphasize that the contracts’ broad language, permitting use for "any lawful government purpose," effectively removes Google’s ability to control how its technology is applied in classified settings.

Broader concerns over military AI use

This is not the first instance of Google employees raising ethical concerns about military collaborations. In April, around 600 employees signed a petition urging leadership to avoid classified military work, reflecting ongoing internal resistance. Similar worries have been voiced about Google’s Project Nimbus, a collaboration with the Israeli military, where critics argue that the company’s ethical guidelines are difficult to enforce in practice.

Google has stated that its contracts include prohibitions on mass surveillance and autonomous weapons, and that it provides channels for employees to raise ethical concerns without fear of retaliation. However, the departing researchers describe these safeguards as weaker in practice than the company claims, particularly in classified environments where transparency is inherently limited.

Pentagon’s push for AI integration

The resignations come as the Pentagon accelerates efforts to integrate advanced AI across classified networks, reflecting a broader trend of militaries seeking to leverage commercial AI technologies. For Google, these contracts represent a significant expansion of its government business, but they also expose tensions between commercial interests and ethical responsibilities.

While Google DeepMind has positioned itself as a leader in AI safety and ethics, the resignations highlight the challenges of maintaining oversight in high-stakes military applications. The departures underscore the difficulty of balancing corporate partnerships with government agencies and the ethical concerns of researchers, who fear unintended consequences in environments where transparency and accountability are constrained.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

For founders and operators in AI, this story is a stark reminder of the ethical and reputational risks that come with government—particularly military—contracts. While such deals can provide revenue and prestige, they also demand clear guardrails and transparency. The resignations at Google DeepMind signal that even well-intentioned policies may not be enough to reassure employees or the public when technology is deployed in classified or high-stakes settings.

Startups working in AI, especially those exploring defense or public-sector collaborations, should consider how they structure oversight, employee dissent, and ethical review mechanisms. The lack of visibility into end-use cases, as seen here, can erode trust internally and externally, potentially leading to talent attrition or public backlash. The story also underscores the importance of aligning commercial partnerships with the values of the team building the technology—values that, for many in AI research, prioritize safety and ethical deployment over unchecked applications.

Related on Lazyfounder

• Google DeepMind
• Pentagon
• More technology news