US Military Leveraged Anthropic’s AI for Iran Air Strike Hours After Trump Order to Halt Use

The United States military, including key commands such as the U.S. Central Command (CENTCOM) operating in the Middle East, reportedly utilized Anthropic’s advanced AI system, Claude, for crucial operational support during a significant air strike against Iran. This operation took place mere hours after President Donald Trump issued a directive to federal agencies mandating an immediate halt to the use of the company’s AI technologies. The Wall Street Journal, citing individuals with knowledge of the matter, detailed that Claude was employed to assist with intelligence analysis, pinpointing potential targets, and conducting battlefield simulations, underscoring the deep integration of sophisticated AI into contemporary defense operations. This incident highlights the complex and often contradictory landscape of AI adoption within the military, where advanced tools remain embedded in critical workflows even as administrative directives seek to sever ties.

The directive from the Trump administration, issued on Friday, specifically instructed government agencies to cease all work with Anthropic and directed the Department of Defense to classify the company as a potential security risk. This sudden policy shift followed a breakdown in contract negotiations, reportedly stemming from Anthropic’s refusal to grant the military unrestricted use of its AI for any lawful scenario as requested by defense officials. The company’s stance, prioritizing ethical boundaries over expansive government contracts, has placed it at the center of a contentious debate regarding the role of AI in military decision-making and operations.

Anthropic had previously secured a substantial, multiyear contract with the Pentagon, valued at up to $200 million, alongside several other leading artificial intelligence laboratories. Through strategic partnerships involving technology giants like Palantir and Amazon Web Services, Claude had been integrated and approved for use in classified intelligence gathering and operational workflows. Reports suggest that the AI system had also played a role in earlier sensitive missions, including a January operation in Venezuela that led to the apprehension of President Nicolás Maduro.

The tensions between Anthropic and the Pentagon escalated significantly when Defense Secretary Pete Hegseth reportedly demanded that the company permit unrestricted military application of its AI models. Anthropic’s CEO, Dario Amodei, rejected this demand, articulating that certain applications represented ethical red lines the company would not cross, even at the cost of government business. In response to Anthropic’s firm stance, the Pentagon began to actively seek alternative AI providers, subsequently reaching an agreement with OpenAI to deploy its AI models on classified military networks. This move signals the Pentagon’s intent to maintain its AI capabilities by diversifying its partnerships.

In a notable interview on Saturday, Anthropic CEO Dario Amodei elaborated on the company’s position, stating that they unequivocally oppose the use of their AI models for mass domestic surveillance and fully autonomous weapons systems. His comments came in direct response to the U.S. government’s directive that labeled Anthropic a defense "supply chain risk" and barred contractors from utilizing its products. Amodei emphasized that specific military applications breach fundamental ethical boundaries and stressed the paramount importance of retaining human control over critical military decisions, rather than delegating them entirely to artificial intelligence.

The situation underscores a broader challenge faced by governments and military organizations worldwide as they grapple with the rapid advancement of AI. The potential benefits of AI in terms of efficiency, intelligence analysis, and operational effectiveness are undeniable. However, these benefits are increasingly juxtaposed with significant ethical, security, and control concerns. The incident involving Anthropic and the U.S. military illustrates the delicate balancing act between leveraging cutting-edge technology for national security and adhering to principles of responsible AI development and deployment.

The breakdown in negotiations between Anthropic and the Pentagon highlights a growing divergence in perspectives regarding the scope and limitations of AI in warfare and intelligence. While military objectives often prioritize the attainment of strategic advantages through any available means, AI developers, particularly those with a stated commitment to ethical AI, are increasingly setting boundaries on how their technologies can be used. This clash of priorities raises critical questions about the future of defense contracting, the governance of AI in military contexts, and the potential for ethical considerations to shape technological adoption.

The Pentagon’s swift move to secure an agreement with OpenAI following the issues with Anthropic demonstrates the urgency with which the U.S. military is pursuing AI integration. OpenAI, another leading AI research laboratory, has also faced scrutiny regarding its work with the military, particularly concerning the potential for its powerful AI models to be used in sensitive or ethically challenging applications. The decision to deploy OpenAI’s models on classified military networks suggests a calculated risk, with the expectation that OpenAI, like Anthropic, will eventually establish its own operational boundaries.

The dual use nature of advanced AI technologies presents a complex dilemma for policymakers and defense leaders. The same AI systems that can be used to identify potential threats and optimize logistics can also be repurposed for surveillance, autonomous targeting, or even cyber warfare. This inherent duality necessitates robust oversight, clear ethical guidelines, and transparent communication between technology providers and government users. The Anthropic case serves as a potent reminder of the need for ongoing dialogue and collaboration to ensure that AI is developed and deployed in a manner that aligns with democratic values and international norms.

The reported use of Anthropic’s Claude AI in a major air strike against Iran, immediately following a directive to halt its use, also raises questions about the practicalities of implementing such directives within complex military operations. The deep integration of AI tools into existing workflows means that immediate disengagement can be technically challenging and operationally disruptive. This highlights the importance of foresight and careful planning when making decisions about the adoption and potential restriction of advanced technologies within defense establishments.

The broader implications of this incident extend beyond the immediate contractual dispute. It signals a potential shift in how the U.S. military approaches its reliance on external AI providers. As AI capabilities become increasingly vital for maintaining a strategic edge, the government’s ability to exert control over the ethical and operational parameters of these technologies will be a critical determinant of future success and public trust. The debate ignited by Anthropic’s stance and the Pentagon’s response is likely to inform ongoing discussions about AI governance, military ethics, and the evolving relationship between Silicon Valley and the defense industrial complex. The future of AI in defense will undoubtedly be shaped by such complex interactions, demanding a careful and continuous recalibration of technological ambition and ethical responsibility.