Pete Hegseth, the U.S. secretary of war, has demanded that Anthropic, a leading AI company, grant the Pentagon unrestricted access to its technology, sparking a national security dispute with Silicon Valley. The confrontation, which has drawn sharp criticism from AI ethicists and tech leaders, highlights growing tensions between the military and the private sector over the control and use of artificial intelligence.

Conflict Over AI Access

Hegseth summoned Dario Amodei, CEO of Anthropic, to the Pentagon last week to demand that the company allow the military to use its AI model, Claude, without restrictions, as long as the use was “lawful.” The move came amid broader concerns about the potential misuse of AI in military applications, particularly in surveillance and autonomous weapons.

Amodei, who has long advocated for the ethical development of AI, has been vocal about the risks of allowing powerful AI systems to be used for military purposes. He has emphasized the need for human oversight and the dangers of fully autonomous weapons, warning that such technology could lead to an “unbeatable army” capable of suppressing dissent within a country.

Amodei’s concerns are not new. In a January essay titled “The Adolescence of Technology,” he wrote that fully powerful AI may be just one to two years away and could surpass human capabilities in fields like biology, programming, and engineering. He also highlighted the psychological impact of AI on individuals, noting the rise in AI-related mental health issues and the potential for mass surveillance.

Hegseth’s demands were met with resistance from Anthropic, which has been reluctant to allow its AI model to be used for surveillance or autonomous weapons without human oversight. The company has stated that it would not comply with the Pentagon’s demands, and on Friday, it announced that it would sue the government over the issue.

Trump’s Role in the Dispute

President Donald Trump has also weighed in on the dispute, criticizing Anthropic and its employees as “Leftwing nut jobs” for helping the Pentagon in its efforts to capture Venezuelan President Nicolás Maduro. Trump directed all federal agencies to stop using Anthropic’s technology, claiming that the company was trying to “STRONG-ARM” the military into following its terms of service instead of the Constitution.

Trump’s comments have raised questions about the administration’s approach to AI regulation. His administration has been accused of having a “healthy disregard for democracy,” particularly in its attempts to undermine election integrity and control the narrative around AI development.

Earlier this year, Hegseth revoked the press passes of journalists who did not agree to sign a pledge restricting their reporting on certain topics. This move, along with Trump’s recent actions, has led to concerns about the administration’s growing authoritarian tendencies.

Hegseth’s ultimatum to Anthropic was to either allow the Pentagon to use its AI model without restrictions or face a designation as a “supply-chain risk,” which would jeopardize the company’s government contracts. This ultimatum put Anthropic in a difficult position, forcing it to choose between compliance and potential legal action.

OpenAI’s Agreement with the Pentagon

Sam Altman, CEO of OpenAI, announced on Friday that his company had reached an agreement with the “Department of War” to use its AI technology on classified work. The terms of the agreement included red lines that appeared similar to those that Amodei sought for Anthropic. Altman praised the Pentagon for its “deep respect for safety” and willingness to collaborate on AI development.

The contrast between OpenAI’s agreement and Anthropic’s refusal has raised questions about the administration’s motivations. Some analysts believe that the administration may have an ideological bias against Anthropic, which is seen as more “woke” than OpenAI. Others speculate that Altman’s personal relationship with Trump may have played a role in securing the deal.

Amodei, however, has not backed down. In a statement, he said that Anthropic would sue the government over the issue, calling the administration’s actions “fundamentally incompatible with American principles.” He emphasized that Anthropic has principles and that the administration is the one that is not aligned with American values.

The dispute highlights the growing tension between the military and the private sector over the control of AI technology. As AI continues to evolve, the question of who should regulate its use and for what purposes remains a pressing concern.

The situation has also drawn attention to the broader implications of AI in national security. Experts warn that if left unchecked, AI could be used for mass surveillance, autonomous weapons, and other applications that could undermine democratic institutions. The conflict between the Trump administration and Anthropic is just one example of the challenges that lie ahead in the regulation of AI.