Home Bots & Bullets‘Pentagon Reviews Ties with Anthropic Over AI Safeguards Dispute’

‘Pentagon Reviews Ties with Anthropic Over AI Safeguards Dispute’

by Pieter Werner

The U.S. Department of Defense is considering ending its relationship with artificial intelligence developer Anthropic amid a dispute over restrictions on how the company’s technology can be used by the military, according to reports by Axios and Reuters.

The disagreement centers on usage limitations Anthropic places on its AI models, particularly its Claude family of large language models. Anthropic has maintained that certain applications — including fully autonomous weapons and mass domestic surveillance — should be prohibited. The Pentagon, in contrast, has pressed for assurances that AI systems can be employed for what it describes as “all lawful purposes,” including weapons development, intelligence collection, and battlefield operations.

Defense officials have expressed concern that Anthropic’s ethical safeguards could make it difficult to deploy the technology in all military contexts, especially in classified environments where flexibility is deemed essential. Other major AI companies such as OpenAI, Google and xAI have agreed to broader terms for unclassified use, but tensions persist over deployment in sensitive defense systems.

One possible outcome is to designate Anthropic as a “supply chain risk,” a classification that would require companies doing business with the Pentagon to avoid using its AI tools. Such a label is typically reserved for foreign adversaries, underscoring the severity of the dispute and its potential impact on Anthropic’s defense partnerships.

Anthropic’s Claude model has been integrated into some U.S. military systems through partnerships with contractors, and is currently among the few commercial AI systems accessible in classified networks. Ending the relationship could affect an existing contract valued at up to about $200 million and shift defense AI strategies toward other providers.

An Anthropic spokesperson said discussions with the Defense Department remained ongoing, focusing on aligning ethical safeguards with operational needs. Pentagon officials did not directly comment on the reports.

Misschien vind je deze berichten ook interessant