Big Tech workers press bosses to back Anthropic in Pentagon clash

Amazon, Google and Microsoft staff are urging executives to back Anthropic in its escalating dispute with the Pentagon, pressing them to refuse any contracts that would enable autonomous weapons or mass domestic surveillance.

In a letter on Friday seen by the FT, worker groups representing thousands of tech employees said they would oppose any effort to dilute guardrails adopted by the AI start-up after its chief executive Dario Amodei rejected what he described as a “final offer” to continue supplying the US military.

“We know [the Pentagon] will rapidly seek to onboard other models without these guardrails in place, regardless of whether they try to force Anthropic to comply,” the letter reads.

“We are writing to urge our own companies to also refuse to comply should they or the frontier labs they invest in enter into further contracts with the Pentagon,” the letter said.

The intervention widens a stand-off that has prompted US defence secretary Pete Hegseth to threaten to rip up Anthropic’s contracts and consider cutting it out of defence supply chains unless it backs down, raising the prospect that rivals could step in.

Emil Michael, under-secretary of defence for research and engineering, on Friday suggested a deal to defuse the situation was still possible.

“I’m open to more talks and I told them so,” Michael told Bloomberg TV, claiming the Pentagon had already made a proposal with “a lot of concessions to the language that Anthropic wanted”. He said that Hegseth would make a decision later on Friday.

OpenAI’s chief Sam Altman told staff on Thursday night that he was working to broker a peace between rival Anthropic and the Pentagon, according to two people familiar with the matter, as first reported by the Wall Street Journal.

The dispute is triggering a broader revolt across Silicon Valley, with rank-and-file engineers warning that the use of autonomous AI for deadly military operations could cross a red line.

“Executive leadership at Google, Microsoft and Amazon must reject the Pentagon’s advances and provide workers with transparency about contracts with other repressive state agencies,” the letter adds, referring to the Department of Homeland Security and Immigration and Customs Enforcement.

Signatories to the letter include the 700,000-member Communications Workers of America, the Alphabet Workers Union, a group of DeepMind staff in London, Amazon Employees for Climate Justice and the umbrella pressure group No Tech for Apartheid.

Microsoft declined to comment. Google and Amazon did not immediately respond.

Google and Amazon have both made multibillion-dollar investments in Anthropic, and Microsoft struck a $30bn cloud-computing deal with the Claude chatbot maker in November. Microsoft also owns 27 per cent of OpenAI.

Anthropic is at present the only AI group approved to work on classified missions, according to an administration official, who added that Claude remained the best model for military use.

Ejecting Anthropic from government contracts would present a commercial opportunity for rivals. OpenAI, Google and Elon Musk’s xAI, which were each awarded $200mn contracts with the defence department last year, are all in discussions with the Pentagon to expand into classified missions, according to people familiar with the talks.

xAI is nearing an agreement allowing the military to use its Grok model without restrictions. Google has not finalised its classified contract or taken a public stance on domestic surveillance and autonomous weapons.

OpenAI was discussing a deal that would also have technical exemptions on “domestic surveillance and autonomous offensive weapons,” Altman told staff on Thursday.

AI researchers and executives have expressed alarm at the Pentagon’s threats. They are concerned about the precedent of banning Anthropic from military contracts as a supply-chain risk or of co-opting its models.

“As an American citizen, the last thing I want is government using AI for mass surveillance of Americans,” said Boaz Barak, a computer scientist at OpenAI, in a post on X.

Mark Chen, OpenAI’s chief research officer, told the FT the company did not rule out signing Pentagon contracts, but that his teams had held constructive debates on the topic of red lines in the usage of AI. He added that there was a “range of opinions” expressed on what was acceptable.

“We don’t know what the other side represents and . . . we need to have a discussion about it internally,” said Chen. “It’s not a top-down decision.”

Google DeepMind’s chief scientist Jeff Dean wrote on Wednesday: “Mass surveillance violates the Fourth Amendment and has a chilling effect on freedom of expression.” He added that he still backed a 2018 commitment to ban lethal autonomous weapons.

More than 100 Google staff sent a letter to Dean asking him to “do everything in your power to stop any deal which crosses these basic red lines”, two people involved told the FT.

More than 270 staff signed an open petition to take the same stance as Anthropic, joined by more than 60 people who said they work for OpenAI.

Staff at Google’s AI research lab have called on the company at large to support its peers against the administration’s demands and maintain guardrails around their own military contracts.

“I think it is unacceptable to nationalise a lab to force them to provide lethal autonomous weapons and mass surveillance tools,” one person wrote in a post liked by hundreds of staff on DeepMind’s internal chat group seen by the FT.

“As an industry, we should stand strong with Anthropic,” they added. “How can we advocate effectively to ensure our AI principles hold firm against this kind of external pressure?”

In February last year, Google quietly dropped a pledge not to use AI technology for weapons or surveillance from its ethical guidelines. Those were first adopted after a staff revolt in 2018 against its military contracts.

More recently, Google has hardened its stance on dissent, firing 28 staff for protesting against a $1.2bn cloud contract with the Israeli government and military.

“Under a normal administration, signing up to all legal use would be unproblematic,” said a person familiar with Anthropic’s position. “This is not a normal administration, and the power of the tech is so much greater.”