Common Cause and its partners are calling on Congress to use oversight powers to investigate the Department of Defense’s overreach in issuing an ultimatum to AI company Anthropic.
Secretary Hegseth is demanding that Anthropic remove restrictions on its use of its AI models to power domestic mass surveillance and fully autonomous weapons. These are two red lines Anthropic is drawing in the sand.
The full text of our coalition letter is below.
To the Chair, Ranking Members, and Committee Members:
On behalf of Common Cause, The Alliance for Secure AI, Young Americans for Liberty, and members and allies across the country, we urge the committees to consider what the current procurement dispute between the Department of Defense and Anthropic actually represents: whether the Department of Defense can expressly reserve the right to violate the law and the constitutional rights of Americans.
The points of contention seem narrow. The Department of Defense and AI company Anthropic are in a public battle over a $200 million contract. At issue are two red lines that Anthropic has drawn in its standard usage policy. The model will not be used for domestic mass surveillance, and it will not be used to power fully autonomous weapons (systems that fire, target, and kill without humans in the decision-making loop). Secretary Hegseth is pressuring the company to eliminate those boundaries and follow a new policy of using AI models for “all lawful purposes,” in line with a January memo that called for the company to be “free from usage policy constraints that may limit legitimate military applications.”
Hegseth gave Anthropic an ultimatum to comply with the new conditions by February 28 or “face the consequences.” These impacts include being designated as a “supply chain risk,” a label for foreign adversaries, and being forced to adjust their models through the Defense Production Act, a law intended for national emergencies.
In doing so, Secretary Hegseth is implying that Anthropic’s red line is inconsistent with his interpretation of the law. The real question is, why hasn’t he promised not to use AI for mass surveillance or fully autonomous weapons?
The decision to take someone’s life is the most serious act a government can take. The Constitution does not leave decisions solely to the discretion of the executive. The laws of war and decades of military doctrine impose accountability at the moment of fatal decisions, precisely because no government, no matter how well-intentioned, can be trusted to police its very borders. Department of Defense Directive 3000.09 has long required “meaningful human control” over the use of deadly force. Anthropic isn’t inventing a new standard. The United States requires the Department of Defense to honor what it is already obligated to do, and the American people have a right to expect their elected representatives to ask why they are not complying.
The issue of monitoring is also based on the same foundation. The Fourth Amendment’s protections against unreasonable searches apply regardless of the technology used. With advanced AI, surveillance that once required vast amounts of resources can now occur automatically, continuously, and at a scale that surprises us all. If existing laws need to be adapted to account for new technology, that is Congress’s job and should not be decided in contract negotiations.
The Department of Defense does not limit this pressure to humanity. OpenAI, Google, and xAI each signed the deal after agreeing to lift standard safeguards for unclassified military systems. This week, xAI formally agreed to the Department of Defense’s “all lawful purposes” standard for deploying Grok models on sensitive military systems without conditions.
The Department of Defense has made it clear that this is not just a human problem. This conflict aims to “set the tone” for all AI companies negotiating with the military. Message received. All other frontier AI companies are already compliant. The only holdouts at this point are the Antropics, and the Pentagon has given them a waiting period until Friday. An example has been created.
The real risks are significant. A “supply chain risk” designation would force all defense contractors to certify that they are independent of Anthropic, whose technology is embedded in eight out of 10 major U.S. companies. “It’s hard to think of a more strategically unwise move for the U.S. military,” said Dean Ball, a former Trump AI adviser who helped develop the Trump administration’s AI action plan.
The decision here is not about which vendor the Department of Defense prefers. The question is whether the federal government can use Frontier AI to conduct mass surveillance and use deadly force in violation of what current law and the Constitution permit. The answer to this question must be a resounding “no.”
These issues require Congressional oversight. We respectfully request that the Committee take the following actions:
- Summon Director Hegseth and senior officials to testify At both the unclassified and classified levels, the department’s requirements for AI companies for “all lawful purposes,” with a particular focus on domestic surveillance capabilities and the development of autonomous weapons.
- Material requests and communications Research from the Department of Defense and Anthropic, OpenAI, Google, and xAI related to the use of AI for domestic surveillance and autonomous weapons. This includes negotiating terms and usage policy agreements with AI contractors. Internal evaluation of requested functionality. and legal analysis supporting the “all lawful purposes” standard.
- Establish reporting requirements Directs the Department to regularly report to Congress on AI capabilities deployed on classified systems, the usage policies governing their deployment, and mechanisms to ensure compliance with the Fourth Amendment and Department of Defense Directive 3000.09. Congress cannot oversee what it cannot see or fully understand.
The issues raised by this dispute are not simply vendor negotiations. These are constitutional and legal questions that belong to the American people and their elected representatives. Americans should not rely on private business as the last line of defense for their constitutional rights and the rule of law. That is Congress’s job.
We call on the Commission to act accordingly.
Sincerely,
brendan steinhauser
CEO
Alliance for Secure AI
carol evans
Vice President of Policy
Common causes
sean theme
chief operating officer
American youth seeking freedom
