Google has reportedly signed a classified deal with the US Pentagon to provide artificial intelligence models for sensitive government work, marking the latest collaboration between a Silicon Valley tech giant and the military. The agreement, reported by The Information, allows the Pentagon to use Google's AI for any lawful government purpose, placing the company alongside OpenAI and Elon Musk's xAI, which have similar contracts.
Details of the Agreement
The deal enables the Pentagon to leverage Google's AI on classified networks, which are used for mission planning, weapons targeting, and other sensitive operations. In 2025, the Pentagon signed agreements worth up to $200 million each with major AI labs, including Anthropic, OpenAI, and Google. Under the terms, Google must adjust its AI safety settings and filters at the government's request. The contract states that the AI system is not intended for domestic mass surveillance or autonomous weapons without human oversight, but it does not grant Google the right to veto lawful government operational decisions.
Employee Backlash
The agreement has sparked significant employee opposition. On Monday, over 600 Google workers signed an open letter to CEO Sundar Pichai, expressing fears that their work could be used in inhumane or extremely harmful ways. They urged the company to refuse making AI systems available for classified workloads. This echoes past protests, such as in 2018 when thousands of employees opposed Google's involvement in Project Maven, a Pentagon contract analyzing drone surveillance footage, leading Google to not renew the contract.
Ethical Concerns and Policy Changes
Last year, Alphabet lifted a ban on using AI for weapons and surveillance, removing language from its ethical guidelines that promised not to pursue technologies causing overall harm. Google's AI lead, Demis Hassabis, stated that AI is crucial for national security. Some employees questioned the shift, with one asking on an internal board, Are we the baddies? The Pentagon has denied interest in mass surveillance or autonomous lethal weapons but insists on allowing any lawful use of AI. Anthropic faced similar tensions earlier this year after refusing to remove guardrails against autonomous weapons or domestic surveillance.
Google stated it supports government agencies on classified and non-classified projects, emphasizing that providing API access with industry-standard practices is a responsible approach to national security. The Pentagon declined to comment on the matter.
