Refuse to become a "war machine"! Over a hundred Google employees co-sign a letter: demand clear boundaries in U.S. military contracts.

```

The fierce struggle between the Pentagon and the AI startup Anthropic over the boundaries of military technology is causing intense reactions in Silicon Valley.

More than 100 Google AI research employees recently submitted a joint letter to management, demanding the company set clear red lines in its cooperation with the US military, refusing to let its technology be used for mass surveillance or fully autonomous weapon systems without human involvement.

On Thursday, over a hundred Google employees sent a joint letter to Jeff Dean, Chief Scientist at its AI division DeepMind, explicitly opposing the use of its Gemini large model by the US military to surveil US citizens or control autonomous weapons. This action echoes Anthropic’s earlier refusal to authorize the Pentagon for “all legal uses.”

Meanwhile, nearly 50 OpenAI employees and 175 Google employees also published an open letter, criticizing the Pentagon's strategy of pressuring tech companies by creating division, urging companies to “set aside differences and unite.”

This incident has cast uncertainty on Google’s soon-to-be-finalized cooperation agreement with the military. As Anthropic faces the threat of losing a $200 million contract by Friday and possibly being deemed a “supply chain risk,” while Musk's xAI has agreed to military terms, Silicon Valley AI companies are undergoing sharp division between commercial interests and ethical boundaries.

Pentagon Pressure Sparks Silicon Valley Backlash

The Pentagon has recently exerted tremendous pressure on Anthropic, demanding it allow the military to use the Claude model for “all legal uses” in classified systems.

US Secretary of Defense Pete Hegseth specifically requested the use of models “unrestricted by policy.” However, Anthropic CEO Dario Amodei refused to compromise, insisting on two red lines—no mass surveillance and no fully autonomous weapons—saying it was a matter of conscience.

This hardline pressure quickly triggered a chain reaction in other Silicon Valley companies. In their letter, Google employees pleaded with Jeff Dean to “do everything possible to prevent any deals that cross these basic red lines,” expressing their desire to be proud of their own work. The letter mentioned that signatories originally planned to oppose “warrantless surveillance of any world citizen,” but deleted this from the letter to increase chances of their demands being met.

Google Executives' Stance and Internal Moral Struggles

As one of Google’s most influential software engineers, Jeff Dean voiced support for Anthropic’s position. He made it clear on social media this week that he opposes government use of AI to surveil Americans, noting that “mass surveillance violates the Fourth Amendment and has a chilling effect on free speech,” and that surveillance systems are prone to misuse for political or discriminatory purposes.

Google has a complicated history handling employee activism. In 2018, a cooperation plan with the Pentagon sparked large-scale employee protests, ultimately forcing the company to abandon the contract renewal. Thereafter, Google centralized relevant decision-making and, in chasing competitors like OpenAI and Anthropic, relaxed some AI safety procedures.

Yet, earlier this month, more than 800 employees petitioned the company to disclose how its technology supports federal immigration enforcement, highlighting ongoing strong internal ethical scrutiny.

“Multi-supplier” Strategy and Risks of AI Arms Race

Confronted by Anthropic’s resistance, the Pentagon is rapidly seeking alternatives. According to Axios and the New York Times, the military has reached an agreement with xAI, allowing its Grok model into classified systems for “all legal uses.”

At the same time, the Pentagon’s negotiations with Google have entered advanced stages, and discussions with OpenAI continue. The Pentagon has even threatened to invoke the Defense Production Act to forcibly requisition Anthropic’s model, and has demanded defense contractors evaluate their reliance on it.

Potential risks of AI models in military applications are not unfounded. According to Tyler Durden’s disclosure of a war simulation led by King’s College London, in 329 rounds of simulation, top AI models ended up choosing nuclear weapons in 95% of cases. Among them, Anthropic’s Claude displayed sophisticated hawkish behavior, decisively launching strikes when risks escalated to the nuclear domain; other models, like Gemini, even used nuclear weapons in the early stages.

Experts warn that AI’s restraint on “nuclear taboo” is far less than that of humans; in a future with highly compressed military decision time, unrestricted AI application may bring disastrous consequences. This is the core reason tech companies insist on drawing red lines.

Risk Warning and DisclaimerThe market carries risks and investment requires caution. This article does not constitute personal investment advice, nor does it consider the unique investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article suit their particular circumstances. Investment is at your own risk. ```