An artificial intelligence company is standing its ground against Pentagon pressure to remove safety restrictions on its technology for military applications, according to a source close to the situation.
Anthropic, an AI research company, met with Defense Secretary Pete Hegseth on Tuesday to address an ongoing disagreement that has stretched on for months. The company has consistently declined to eliminate protective measures that block its artificial intelligence systems from being used for autonomous weapons targeting or domestic surveillance operations.
The dispute centers on Anthropic’s unwillingness to modify safeguards built into its AI technology. Pentagon representatives have maintained that government agencies should only need to follow existing U.S. legal requirements when using such systems.
According to the source, Hegseth presented Anthropic’s leadership with a stark choice during Tuesday’s discussion: either accept being classified as a supply-chain security risk, or face the government using legal authority to compel changes to the company’s usage policies. The administration has set a Friday deadline for Anthropic’s response.
Anthropic CEO Dario Amodei participated in the high-stakes meeting, but the company remains committed to maintaining its current restrictions, the source indicated.
Pentagon representatives have not yet provided comment on the meeting or the ultimatum.
