Anthropic boss rejects Pentagon demand to drop AI safeguards

Defense Secretary Pete Hegseth previously threatened to remove the firm from the department's supply chain.

Nvidia CEO Jensen Huang has weighed in on the conflict between the Pentagon and AI firm Anthropic, stating the situation is 'not the end of the world.' The dispute centers around Anthropic’s refusal to allow its Claude AI model to be used for all lawful military purposes, including potentially controversial applications such as autonomous weapon systems and mass domestic surveillance. While the Pentagon insists contractors comply with these terms, Anthropic is resisting due to ethical concerns. The standoff could jeopardize Anthropic's $200 million contract with the Department of Defense, and Defense Secretary Pete Hegseth has threatened to invoke the Defense Production Act (DPA) to force compliance, citing the company as a 'supply chain risk.' Huang acknowledged both sides’ viewpoints and emphasized hope for a resolution, though he noted that a failure to resolve the issue wouldn't be catastrophic. Nvidia has a vested interest in the outcome through its $5 billion investment in Anthropic in exchange for adoption of its GPU architecture. The Pentagon has given Anthropic until Friday to comply.

In a high-stakes standoff, U.S. Defense Secretary Pete Hegseth has issued an ultimatum to AI company Anthropic, demanding the removal of safety guardrails from its Claude AI system by Friday or face severe repercussions. These include the cancellation of a $200 million Pentagon contract, classification as a 'supply chain risk,' or enforcement action under the Defense Production Act. The confrontation stems from tensions over Anthropic CEO Dario Amodei's ethical concerns about government misuse of AI, particularly in autonomous weapons and mass surveillance. Despite Anthropic’s involvement in national security efforts, the company has drawn criticism for resisting domestic surveillance and weaponization of AI, which some within the Trump administration have labeled as 'Woke AI.' The latest meeting ended in a stalemate, with Amodei holding firm on the company's principles. Legal experts and commentators have questioned the legitimacy and wisdom of the Pentagon's threats, highlighting that critical military AI policy is being shaped by private negotiations rather than legislative oversight.