The Department of Defense has intensified requests to gain access to Anthropic’s Claude model for military applications, according to media reports. Officials reportedly delivered a final offer and asked for a response by the end of the week.

Terms requested by the Pentagon

According to the reports, the military seeks the ability to deploy the model with minimal limitations, constrained only by existing law and regulation. The proposal would allow broad operational use across defense tasks without additional vendor-imposed restrictions.

Anthropic’s position

Anthropic has insisted on retaining guardrails and oversight, citing concerns about potential AI errors and unpredictable behavior in operational contexts. The company emphasizes that final decisions, particularly during combat operations, should remain under human authority.

Possible implications

Sources indicate that a refusal could lead to the loss of existing Army contracts and trigger a designation as a risk within the defense supply chain. Such a designation might limit future procurement opportunities for the vendor.

Outlook

Negotiations are reportedly time-sensitive, with the Pentagon setting a near-term deadline for a response. Both sides are balancing operational requirements, safety concerns, and legal responsibilities as discussions proceed.