Deadline looms as Anthropic rejects Pentagon demands it remove AI safeguards

Anthropic CEO Dario Amodei rejected the Pentagon's ultimatum to remove safety restrictions from the company's AI model in Washington on February 26, 2026, as the Defense Department threatened to blacklist the company from lucrative military contracts over disagreements on how artificial intelligence should be used in warfare. The conflict centers on Anthropic's refusal to allow its Claude AI model to be used for mass surveillance of Americans or to power fully autonomous weapons, which Amodei has described as 'entirely illegitimate' and 'bright red lines' for the company. Despite the Pentagon's insistence that it doesn't intend to use Anthropic's tools for these purposes, the Defense Department has maintained that it's not up to private companies to decide how their technology is used, requiring AI companies to allow their tools 'for all lawful purposes.' The escalating tensions reached a critical point when Defense Secretary Pete Hegseth threatened to cancel Anthropic's $200 million contract and potentially designate the company as a supply chain risk, an unusual designation typically reserved for foreign adversary technology like Huawei. In response, Amodei issued a statement explaining that while Anthropic is committed to using AI to defend democracies, certain applications 'can undermine, rather than defend, democratic values.' The Pentagon has given Anthropic a hard deadline of 5:01 PM ET on February 27, 2026, to comply with its demands, after which it will terminate the partnership and take unspecified action against the company. This confrontation represents a significant moment in the relationship between AI companies and the U.S. military, particularly as Anthropic was the first AI company cleared for classified use by defense officials, who deemed its model the most advanced and secure for sensitive military applications. The Pentagon has also threatened to invoke the Defense Production Act to force Anthropic to remove its guardrails, an extraordinary step traditionally reserved for true emergency crisis situations. Legal experts predict that if the Pentagon escalates the situation by attempting to compel compliance or implement a supply chain risk designation, Anthropic will likely respond with legal challenges, prolonging what could become a landmark case in determining the boundaries of corporate control over AI applications in national security contexts.