Anthropic, on March 9, filed a lawsuit to block the Pentagon from adding it to a national security blacklist, intensifying the artificial intelligence laboratory’s high-stakes competition with the US military over usage restrictions on its technology.
The Pentagon on March 5 officially supply-chain risk designation on Anthropic, restricting the use of a technology, which a source indicated was being utilized in military operations in Iran.
In its lawsuit, Anthropic claimed that the designation was illegal and infringed upon its freedom of speech and due process rights. The case filed at the federal court in California requested a judge to reverse the designation and prevent federal agencies from enforcing it.
Anthropic said, “These actions are unprecedented and unlawful. The Constitution does not allow the government to wield its enormous power to punish a company for its protected speech.”
Anthropic has been deemed a national security supply-chain risk by the Defence Secretary Pete Hegseth, who made the designation last week following a refusal by the startup to implement guardrails against the use of its AI in autonomous weaponry or domestic surveillance.
Its designation presents a massive risk to the business of Anthropic with the government, and the decision may impact how other AI vendors negotiate against the use of their technology by the military, albeit with the Chief Executive Officer of the company Dario Amodei explaining on March 5 that the designation had “a narrow scope” and businesses could still utilize its tools in projects unrelated to the Pentagon.
President Donald Trump has also instructed the government to cease collaboration with Anthropic, whose funding sources are Alphabet, Google, and Amazon.com. Trump and Hegseth indicated that there would be a 6-month phase-out.
Reuters has also stated that the investors in Anthropic were scrambling to limit the damage inflicted by the blowback with the Pentagon.
Trump and Hegseth acted on February 27, following a series of negotiations with Anthropic on whether policies of the company could limit military action, and shortly after, Amodei sat with Hegseth with hopes of striking a deal.
The Pentagon noted that US legislation, and not a commercial enterprise, would decide on the means of defense for the nation and demanded that it be given full latitude in applying AI for “any lawful use,” claiming that the limitation on the use of AI provided by Anthropic would pose a threat to American lives.
Anthropic reported that even the top AI models were not reliable enough for fully autonomous weapons and that using them for that purpose would be dangerous. The company also drew a red line on domestic surveillance of Americans, calling it a violation of fundamental rights.
Anthropic, in a statement made after the announcement by Hegseth, stated that the name was legally unsound and would provide a bad precedent to other companies that enter into agreements with the government.
The company said it would not be cowed by “intimidation or punishment,” and on March 5, Amodei reiterated that Anthropic would challenge the designation in court.
He also apologized for a memo within his company that was released on March 4 by tech news The Information. In the memo, which was written on February 27, Amodei claimed that Pentagon officials did not like the company in part because “we haven’t given dictator-style praise to Trump.”
The Defence Department signed agreements valued at up to US$200 million (S$255 million) each with massive AI labs in the past year, including Anthropic, OpenAI, and Google.
Microsoft-based OpenAI signed an agreement to implement its technology in the Defense Department network soon after Hegseth listed Anthropic on the blacklist.
Chief executive officer Sam Altman said the Pentagon shared OpenAI’s principles to ensure that human oversight of weapon systems and opposing mass US surveillance.
