Department of War Drops Anthropic Over Safeguards, Turns to OpenAI Instead

Anthropic has reportedly been cut off from U.S. military contracts after refusing to allow its Claude AI models to support certain defense-related uses.

According to Neowin, the dispute escalated after Anthropic objected to potential deployments involving mass domestic surveillance and fully autonomous weapons systems. Federal agencies were then ordered to stop using Anthropic’s AI tools altogether.

Anthropic labeled a supply chain risk

The conflict intensified when Secretary of War Pete Hegseth designated Anthropic as a supply chain risk. That classification would effectively prevent the company from securing military-related contracts.

The Department of War reportedly stated it would only work with AI firms that agree to “any lawful use” of their technology and remove restrictive safeguards that limit deployment scenarios.

Anthropic strongly rejected the move, calling the designation legally unsound and a dangerous precedent for companies negotiating with the federal government. The company said it plans to challenge the decision in court and reaffirmed its opposition to domestic mass surveillance and fully autonomous weapons use.

The standoff highlights rising tension between AI developers who promote strict safety frameworks and government agencies that demand broader operational flexibility.

OpenAI reaches an agreement with the Department of War

While Anthropic faces exclusion, OpenAI has confirmed it reached an agreement with the Department of War.

CEO Sam Altman stated that OpenAI’s models will be deployed inside the department’s classified network. He emphasized that the deal preserves key principles, including prohibitions on domestic mass surveillance and maintaining human responsibility in the use of force, even when autonomous systems are involved.

OpenAI also plans to implement technical safeguards and place engineers within the Department of War environment to monitor model behavior and ensure compliance with agreed standards.

Officials suggested that the OpenAI contract still satisfies the government’s “all lawful use” requirement, referencing existing legal authorities and mutually agreed protections. However, the full terms of the agreement have not been made public.

It remains unclear why Anthropic and the Department of War failed to reach a similar compromise.

Growing divide over military AI policy

The dispute underscores broader friction between AI companies’ ethical policies and U.S. national security priorities. As defense agencies expand interest in AI-driven analytics, automation, and decision support, debates around surveillance boundaries and autonomous weapons continue to intensify.

In parallel, OpenAI recently secured major funding commitments from Amazon and NVIDIA, strengthening its position as a leading supplier of advanced AI systems across both commercial and government sectors.

The unfolding situation signals that AI policy, procurement, and national security will remain closely intertwined as governments and private companies negotiate the limits of artificial intelligence in military contexts.