Pentagon stops using Anthropic AI tools after blacklisting company, BBC told

The United States Department of Defence has ceased using Anthropic's AI tools, according to an official statement released on Monday. This decision comes months after the Pentagon designated Anthropic as a supply chain risk on national security grounds. Secretary of Defence Pete Hegseth had labeled Anthropic as a risk in February, giving the company until late August to remove its products from Pentagon usage. However, there has been speculation surrounding the delay in completing this transition.

Multiple sources familiar with the matter revealed that Anthropic's Claude model was still being utilized by the department in research, analysis, intelligence gathering, and military operations against Iran. This powerful tool was instrumental in sorting through vast amounts of information and aiding in the identification of potential military targets. Former US defense officials and contractors working closely with the Pentagon confirmed that Claude had been a critical component in their operations.

Anthropic's Claude model was particularly valuable in the Maven Smart System, a primary data platform operated by Palantir, which is the Pentagon's main platform for organizing intelligence and other data. It is believed that cleaning up Claude's integration into Maven took time and effort, making its removal a challenging process.

The Pentagon's decision to stop using Anthropic's tools took place after the company refused to remove Claude's safety guardrails and grant the military unfettered access to its AI products. Anthropic argued that doing so would result in mass surveillance and the creation of autonomous weapons, which they deemed both unethical and unlawful. The company has even taken Anthropic to court to challenge the designation, describing it as "unprecedented and unlawful."

During his tenure as Secretary of Defence, Hegseth gave Anthropic a six-month deadline to phase out of government use on February 27, 2024. This deadline occurred just one day before the US and Israel attacked Iran. Despite the controversy surrounding Anthropic, other major tech companies, such as Google, xAI, and OpenAI, have been able to secure contracts with the Department of Defence more quickly.

The relationship between the Pentagon and the Trump administration appears to be warming, with Trump meeting Anthropic's CEO Dario Amodei twice at the White House. During one of these meetings, Amodei expressed concern about AI risks, proposing a cooperative approach to address them. However, Trump subsequently praised Amodei, describing him as "great" and "fantastic."