AI Governance Clash: Pentagon and Anthropic Face Off Over Military AI Use

A dispute between the United States Department of Defense (DOD) and Anthropic has escalated, raising a critical question: who should set the guardrails for military use of artificial intelligence – the executive branch, private companies, or Congress and the broader democratic process?

The Standoff Begins

The conflict began when Defense Secretary Pete Hegseth reportedly requested that Anthropic CEO Dario Amodei allow the DOD unrestricted use of its AI systems. When Anthropic refused, the administration moved to designate the company a “supply chain risk” and ordered federal agencies to phase out its technology, significantly escalating the situation. President Trump subsequently ordered all US agencies to stop using Anthropic’s AI.

Anthropic’s Red Lines

Anthropic has drawn two key lines: preventing its models from being used for domestic surveillance of United States citizens and prohibiting fully autonomous military targeting. Secretary Hegseth has objected to what he describes as “ideological constraints” embedded in commercial AI systems, arguing that the government, not vendors, should determine lawful military use.

A Procurement Disagreement?

At its core, this dispute resembles a procurement disagreement. The U.S. Military decides what it wants to buy, and companies decide what they are willing to sell and under what conditions. However, the decision to designate Anthropic a “supply chain risk” shifts the dynamic from a standard procurement issue to the use of coercive leverage, potentially impacting any contractor doing business with the U.S. Military. Anthropic’s CEO, Dario Amodei, has stated he would challenge this designation in court.

Civil Liberties and Autonomous Weapons

The issues at stake touch on fundamental concerns. Anthropic’s opposition to domestic surveillance aligns with established civil liberties protections. The company argues that it should not be required to build systems that could be used to violate those rights, even if the government claims lawful authority. The disagreement centers on whether these constraints should be imposed by law and oversight, or by developers through technical design.

The question of fully autonomous military targeting is more complex. Although the DOD already has policies requiring human judgment in the use of force, the military may see such capabilities as necessary for deterrence and operational effectiveness. Anthropic reasonably believes its current technology is not sufficiently reliable or controllable for certain battlefield applications.

The Need for Clear Governance

The boundaries of military AI use should not be settled through negotiations between a Cabinet secretary and a CEO. If the U.S. Government believes certain AI capabilities are essential, that position should be openly debated in Congress and reflected in doctrine, oversight mechanisms, and statutory frameworks. Clear rules are needed, not only for companies but for the public.

Strategic Implications

If companies conclude that participating in federal markets requires surrendering all deployment conditions, some may exit those markets. Others may weaken safeguards to remain eligible for contracts. Neither outcome strengthens U.S. Technological leadership. Built-in safeguards can complement existing oversight structures, reducing the risk of misuse or unintended escalation.

Congressional Role

Congress should clarify statutory boundaries for military AI use and investigate whether sufficient oversight exists. The DOD should articulate detailed doctrine for human control, auditing, and accountability. Civil society and industry should participate in structured consultations. If AI guardrails can be removed through contract pressure, they will be treated as negotiable; if grounded in law, they can become stable expectations.

This episode highlights a pivotal moment in AI governance. AI systems are now powerful enough to influence intelligence analysis, logistics, cyber operations, and potentially battlefield decision-making, making them too consequential to be governed solely by corporate policy or executive discretion.

This article is adapted with permission from Tech Policy Press.