In the clearest and most consequential policy move yet on AI safety, the Trump administration announced it would blacklist major AI labs for denying them unfettered access to the technology for military purposes.
The president and Army Secretary Pete Hegseth are the ones who are aiming for the nuclear issue. Anthropic denies Department of Defense use of company AI for ‘any lawful purpose’.
Describing Anthropic as a woke radical left-wing company, the president said on his Truth Social platform that “the left-wing lunatics at Anthropic made a disastrous mistake in trying to strengthen the Department of the Army,” adding that the company’s actions are putting American lives and national security at risk.
But to date, Anthropic has done more to support the Department of Defense than any other AI institute.
Anthropic’s Claude AI is the only Frontier model already widely used in sensitive military planning and operations.
Claude AI was widely reported to have been used as part of the Pentagon’s “Maven smart system” to plan and execute a military operation to capture Venezuelan President Nicolas Maduro in January.
The controversy was not about Anthropic’s involvement with the U.S. military. Instead, it advocated “red lines” regarding the use of AI technology.
Anthropic CEO Dario Amodei called for guarantees that it would not be used for mass surveillance of civilians or deadly automated attacks without human oversight.
Anthropic CEO Dario Amodei is at odds with the Pentagon. Photo: Reuters
Amodei said in a statement Wednesday that some uses of AI are “completely beyond what can be done safely and reliably with today’s technology.”
In a similarly impassioned post to X, Secretary Hegseth announced that Anthropic would not only be placed on the blacklist, but would also be designated as a supply chain risk. This is a legal intervention previously limited to foreign technology companies deemed to be a direct threat to U.S. national security.
read more:
Rapid developments in AI are ‘making it harder to measure’‘
The AI bubble remains intact for now.
It’s a move that shocked AI safety activists, given growing concerns about AI safety, but it also raises serious questions about the future viability of the Pentagon’s “AI first” strategy.
Secretary Hegseth gave Anthropic six months to remove its AI from Department of Defense systems. But now the question is what he will replace it with.
For the first time in the short history of superintelligent AI, this uproar appears to have united the AI industry.
OpenAI CEO Sam Altman, who is also in talks with the Pentagon, announced in a memo to staff Thursday that the company shares the same “red lines” as Anthropic.
Separately, more than 400 Google and OpenAI employees signed an open letter calling for the entire industry to come together to oppose the Department of the Army’s position.
OpenAI boss Sam Altman has concerns about the use of AI technology in warfare. Photo: Reuters
In a copy of the OpenAI memo seen by Sky News, Altman told staff: “Regardless of how we got here, this is no longer just an issue between Anthropic and DoW. This is an industry-wide issue and it’s important that we make our position clear.”
The Trump administration’s move thus appears to be as much about power as it is about AI safety.
The Pentagon has already said it will not use AI or unsupervised autonomous weapons for mass surveillance of American citizens.
Photo: Reuters
The company’s furious response to Anthropic seems more like a reaction to big tech companies trying to dictate terms to the government than what those terms actually are.
In taking on Silicon Valley, the administration has just declared war on a powerful adversary, even though AI investment accounts for much of current U.S. economic growth.
