Google has entered a classified AI agreement with the Pentagon, allowing its models to be used for “any lawful government purpose.” This move comes at the same time as significant internal dissent, with more than 600 employees urging leadership to avoid military applications of AI.
Key developments:
- Over 600 employees at Google sent a letter to CEO Sundar Pichai, asking him to “refuse to make our AI systems available for classified workloads.”
- Reporting from The Information indicates the deal permits use of Google’s AI for “any lawful government purpose,” without granting the company legal authority to restrict how the Pentagon deploys it.
- Rival firms, including OpenAI and xAI, finalized similar Pentagon agreements last month. Meanwhile, Anthropic is reportedly engaged in legal challenges after being excluded for refusing to relax its safety guardrails.
- Google’s earlier commitment to avoid weapons, related AI use, introduced in 2018 after employee protests, was removed from its AI principles in 2025.
Why important?
This situation underscores mounting tension in the AI sector between national security partnerships and internal ethical concerns. Although recent disputes, particularly involving OpenAI and Anthropic, have kept Pentagon collaborations in the spotlight, they have not slowed broader industry participation. Google now faces both reputational and internal pressure as it moves deeper into defense-related AI work, raising the possibility of further backlash as its systems, including Gemini, gain wider adoption.
Sources:
https://www.theguardian.com/technology/2026/apr/28/google-classified-ai-deal-pentagon
https://www.businessinsider.com/google-employees-ceo-block-classified-military-ai-projects-2026-4
https://www.cbsnews.com/news/google-ai-pentagon-classified-use-employee-letter