Anthropic Stands Firm Against Pentagon’s AI Weapons Demands

TLDR

• Anthropic faces Pentagon pressure to eliminate safety restrictions on Claude AI for unrestricted military applications, including autonomous weaponry and surveillance operations.
• CEO Dario Amodei declined the request, citing risks to democratic principles.
• Military officials established a Friday 5pm ultimatum for compliance or exclusion from defense agreements.
• Defense officials warned of potential Defense Production Act enforcement and classification as a security threat to the supply chain.
• Latest contract revisions delivered Wednesday evening were deemed inadequate by Anthropic.

Dario Amodei, leading Anthropic as CEO, has maintained his position against removing protective measures from the Claude AI system, even as this stance threatens a significant federal partnership. Military officials have imposed a Friday cutoff time, insisting the company must consent to “any lawful use” of its AI platform.

The core disagreement revolves around two particular applications: deploying Claude for large-scale domestic monitoring operations and enabling completely autonomous weapon systems. According to Anthropic, neither application was ever included in their existing Pentagon arrangements and shouldn’t be introduced at this stage.

Amodei engaged in discussions with Defense Secretary Pete Hegseth during the current week. Those talks concluded without resolution, prompting the Pentagon to submit updated contractual terms on Wednesday evening.

Military leadership has been direct with its warnings. Officials indicated they would exclude Anthropic from military partnerships and classify the organization as a “supply chain risk” — a categorization usually applied to vendors from adversarial countries.

A high-ranking Pentagon source also informed Reuters that Secretary Hegseth might utilize the Defense Production Act. This legislation enables the government to compel corporate participation in national security initiatives, regardless of company agreement. Legal scholars have raised doubts about whether such application of the statute would be constitutional.

What Anthropic Says About AI Weapons and Surveillance

Anthropic expressed support for AI applications in legitimate foreign intelligence operations, while opposing domestic surveillance programs.

The Business Risk for Anthropic

The monetary implications are substantial. Over the previous twelve months, the Pentagon has established $200 million framework agreements with prominent AI companies including Anthropic, OpenAI, and Google.

Should the company receive a supply chain risk designation, military contractors such as Lockheed Martin would be prohibited from utilizing Anthropic’s technology on Department of Defense initiatives. The defense contractor ecosystem encompasses approximately 60,000 companies.

Amodei indicated that Anthropic proposed collaborating with military officials on research and development efforts to enhance AI dependability for defense applications, but this proposal was declined.

As of Thursday evening, both parties remained deadlocked with the 5:01 p.m. Friday time limit unchanged.

The post Anthropic Stands Firm Against Pentagon’s AI Weapons Demands appeared first on Blockonomi.