Google signs classified AI contract with Pentagon, sparking employee backlash
Key Points
- Google signed a classified contract with the U.S. Department of Defense allowing use of its AI models for any lawful government purpose.
- The deal was reported by The Information on Tuesday, citing an anonymous source.
- More than 560 Google employees signed an open letter on Monday urging CEO Sundar Pichai to refuse classified military AI work.
- Unlike Anthropic’s earlier Pentagon agreement, Google’s contract contains no ethical carve‑outs on surveillance or autonomous weapons.
- OpenAI and xAI have already signed similar classified AI contracts with the Pentagon.
- Anthropic was blacklisted earlier this year after refusing to remove ethical restrictions from its contract.
- Google has not publicly confirmed the terms of the agreement, and the company’s response to employee concerns remains pending.
Google has entered a classified agreement with the U.S. Department of Defense that lets the Pentagon use the company’s artificial‑intelligence models for any lawful government purpose. The deal, reported by The Information, comes just a day after more than 560 Google engineers signed an open letter urging CEO Sundar Pichai to refuse such military contracts. Unlike an earlier Anthropic pact that barred mass surveillance and fully autonomous weapons, Google’s terms contain no ethical carve‑outs. The move adds the tech giant to a short list of AI firms supplying unrestricted AI capability to the U.S. military.
Google announced a classified partnership with the U.S. Department of Defense that grants the Pentagon unrestricted access to the company’s artificial‑intelligence models. The agreement, described by The Information as allowing use for “any lawful government purpose,” marks the latest addition to a growing roster of AI vendors supplying the military with advanced capabilities.
Employees at Google reacted sharply. On Monday, more than 560 engineers and staff signed an open letter addressed to CEO Sundar Pichai, urging him to reject any classified military AI work. By Tuesday, the same company had sealed a deal the signatories had asked it to refuse, creating a stark contrast that will likely dominate internal town halls and public statements.
The contract differs markedly from the arrangement Anthropic struck with the Pentagon earlier this year. Anthropic’s deal included explicit prohibitions on mass domestic surveillance and fully autonomous weapons without human oversight. Those restrictions led the Trump administration to blacklist Anthropic as a national‑security supply‑chain risk in February 2026. Google’s agreement, by contrast, lacks such ethical limits, aligning with the broader, less‑restricted model favored by the administration.
OpenAI and Elon Musk’s xAI have already signed similar classified contracts. OpenAI renegotiated its terms to retain some red lines on domestic surveillance, while xAI reportedly entered the agreement without notable restrictions. Google now joins a quartet of AI firms—OpenAI, xAI, Google, and formerly Anthropic—capable of providing the Pentagon with cutting‑edge AI tools for classified missions.
The timing of the deal amplifies the internal tension. The employee letter was circulated on Monday morning; the classified contract was reportedly signed by the next day. Analysts expect the episode to surface in upcoming Pichai town halls, press briefings, and possibly even legal testimony in the high‑profile Musk‑Altman trial, where questions about corporate AI ethics could arise.
Google has not publicly confirmed the specific terms of the Pentagon engagement. The “any lawful government purpose” language comes from a single anonymous source cited by The Information. Without official comment, the company’s stance remains unclear, leaving employees and external observers to interpret the implications.
The broader debate pits government demand for unrestricted AI against corporate commitments to ethical AI development. Since the 2018 Project Maven controversy, many tech firms pledged to avoid weaponizing AI without human oversight. Anthropic’s blacklisting demonstrated the cost of maintaining those principles, while OpenAI and now Google appear willing to accommodate the Pentagon’s broad needs.
Whether Google’s decision proves temporary or permanent will hinge on political shifts and internal pressure from the signatories of the open letter. For now, the deal expands the Pentagon’s pool of AI suppliers, granting it access to some of the most powerful models on the market while leaving ethical safeguards largely undefined.