Pentagon Halts Use of Anthropic's Artificial Intelligence Tools

The Pentagon has halted the use of artificial intelligence tools developed by Anthropic. The suspension came after the agency placed the company on a restrictive list, according to information provided to the BBC.

The decision to label Anthropic a supply‑chain risk was made in February. That determination followed the company's refusal to remove safety safeguards built into its technology solutions.

Safeguards and Operational Control

Safety safeguards in AI models consist of technical guidelines and parameters inserted by developers to limit certain behaviors and restrict unwanted outputs from the systems. These mechanisms work to prevent the tools from taking actions that violate the rules set out in their original design.

Designating an organization as a supply‑chain risk functions, in general public‑administration terms, as an institutional tool to block or discontinue the acquisition and use of technologies from specific vendors. Inclusion in this category prevents the integration of a supplier's products or services into the agency's infrastructure.

In the Pentagon's case, the developers' maintenance of these protection barriers and the formal classification in February led to the end of the company's resources being used in its operations.

With information from BBC News.