The Silicon Valley commander-in-chief?

Who holds the final authority when American lives are at stake? Is it the commander-in-chief or a Silicon Valley CEO? A growing standoff between the Pentagon and the AI startup Anthropic has moved this question from science fiction into the halls of power. At the heart of the dispute is a fundamental question of governance. Can a private corporation use its own internal ethics protocols to veto the national security decisions of a sovereign government?

The dispute between Anthropic and the United States Department of War is one of the first major political conflicts over how advanced AI should be governed in national security. Anthropic’s large language model (Claude) had already been integrated into U.S. defense and intelligence systems through a lucrative $200 million contract with the Pentagon. This allowed the military to use Claude for operational analysis and intelligence support in missions such as the U.S. raid that captured Nicolás Maduro.

However, Anthropic has asserted strict internal red lines tied to the company’s ethics protocols. These restrictions created friction with defense officials. When Pentagon leaders pushed for broader operational flexibility and asked to use Claude for all lawful purposes, Anthropic refused to remove its restrictions and argued that it needed assurances its technology would not be used in ways the company considered unethical.

The standoff escalated quickly. In a meeting with Pentagon officials, a hypothetical scenario was raised. If the United States were under attack and hypersonic missiles were heading toward U.S. territory, and Claude could help stop them, would Anthropic block that use because of its ethical policies? Anthropic’s CEO Dario Amodei reportedly said the military should check with Anthropic for clearance even during such an attack. Pentagon officials viewed this response as proof that relying on the company’s AI in a crisis could introduce dangerous uncertainty.

As reported by the Wall Street Journal, this exchange exemplified the different worldviews between Secretary of War Pete Hegseth and Amodei. Unwilling to allow a private company executive to hold what amounts to veto power over military operations or the commander-in-chief, Hegseth ultimately moved to designate Anthropic as a supply chain risk. This decision cut the company off from federal contracts worth hundreds of millions of dollars.

Such a designation carries serious consequences. The Pentagon can bar Anthropic from procurement programs, prevent other defense contractors from integrating Anthropic technology into military systems, and steer federal spending toward alternative suppliers. For a frontier AI company this could mean losing hundreds of millions of dollars in potential government work as well as access to defense data, partnerships, and strategic collaboration. The designation has rarely been applied to an American company and there is a legitimate debate about whether it is an appropriate response by the Department of War.

Yet the larger significance of this episode goes beyond one contract dispute. It represents a defining test of who ultimately governs advanced AI in wartime. The question is whether private companies working as government contractors will determine how these tools may be used, or whether the U.S. government will assert authority over technologies it considers essential to national defense.

In the aftermath of this fight, Anthropic has seemingly stepped back from its initial concerns about Claude being used in the Maduro raid. Instead, the company has emphasized a new red line, that it should not be used for surveillance or for autonomous weapons systems. At first glance this concern sounds reasonable. AI could absolutely enable new forms of surveillance that many Americans would find troubling.

It is obvious that the Fourth Amendment is an essential safeguard of our republic. But questions about whether intelligence agencies such as the NSA may collect certain information are ultimately questions for Congress and the courts. The surveillance scenario assumes all surveillance is inherently illegal. In reality, governments routinely conduct lawful intelligence collection under statutory authority and judicial oversight.

If AI were being used for illegal purposes, I would be the first to argue that Anthropic should refuse to participate. But that is not the situation here. The question of whether AI might enable new forms of surveillance that raise constitutional concerns is a serious one. Yet those questions should be debated through democratic institutions rather than determined by the internal policies of a single company executive.

If Anthropic truly believes AI will transform surveillance capabilities, then the company should lead the effort to rethink authorities such as Section 215 and Section 702 of the Foreign Intelligence Surveillance Act (FISA). Those are the forums where these issues belong. Yet a quick look at Anthropic’s lobbying disclosures or public policy statements suggests the company has not engaged in that debate.

The same logic applies to the question of autonomous weapons. It is easy to understand why people are uneasy about the use of these emerging technologies. Those concerns deserve careful discussion. But decisions about national defense should not be determined by a single government contractor. They should be debated in the public square and resolved through democratic governance.

This is not an easy or simple issue. Anthropic is raising real ethical questions about powerful new technologies that policymakers and citizens alike should grapple with. At the same time, it is understandable why the Department of War is unwilling to allow a private CEO to act as a de facto commander-in-chief—effectively involved directly in battle decisions.

As Christians, it is essential that we engage in this debate. Questions of security and technology are areas where moral reflection matters deeply. AI will shape warfare, our economy, and our culture in ways we are only beginning to understand.

That moral reflection is necessary. Yet it must operate within the institutions God has ordained for civil authority. In the American system, the use of force and the defense of the nation are entrusted to the state, accountable to the people through law and democratic governance.