Anthropic CEO Dario Amodei uncertain about AI model’s role in Iran missile strike

Anthropic CEO Dario Amodei admitted uncertainty over whether the company’s AI model, Claude, was used in the recent US-Israeli missile strikes against Iran. The tension stems from a Pentagon ultimatum demanding unrestricted military access to Claude, which Anthropic refused, leading to a ban and labeling the company a supply-chain risk.
Anthropic CEO Dario Amodei has expressed uncertainty about whether the company’s AI model, Claude, played a role in the US and Israeli missile strikes against Iran in early March 2026. The admission comes amid a heated dispute between Anthropic and the Pentagon, which escalated after Defense Secretary Pete Hegseth demanded unrestricted access to Claude on February 24, 2026. Claude had already been integrated into US military intelligence workflows for target identification when the Pentagon issued its ultimatum. Two days later, Anthropic publicly rejected the Pentagon’s demands, stating it would exclude mass domestic surveillance and fully autonomous weapons systems from defense contracts. The Pentagon responded by banning Anthropic’s technology from defense systems and labeling it a supply-chain risk. The strikes on Iran targeted around 1,000 locations within the first 24 hours, a scale that suggests AI-assisted intelligence workflows may have been involved. Despite the subsequent ban, there is a possibility that military operators retained access to Claude’s capabilities or legacy data processed through its models during the planning or execution of the strikes. Amodei clarified that current frontier AI models are not capable of fully autonomous weapons deployment but can assist humans in decision-making, such as target prioritization. The situation highlights the deep integration of AI into national security infrastructure, with Anthropic’s refusal to grant unrestricted access marking a clear fracture in its relationship with the Pentagon. Founded by former OpenAI researchers, Anthropic has positioned itself as a leader in responsible AI development. However, its prior collaboration with the military—including embedding Claude in targeting systems—has created a conflict between its safety-first mission and military demands. The fallout from this dispute raises broader questions about AI’s role in modern warfare and the ethical boundaries of its use.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.