Senate Democrats Push to Enshrine Anthropic's AI Guardrails on Weapons, Surveillance.

Senator Adam Schiff has expressed a clear stance against solely relying on the assurances of the Pentagon or AI company executives concerning the deployment of autonomous weapons.

The ongoing dispute between AI firm Anthropic and the Pentagon is now escalating to Congress. Senator Adam Schiff (D-CA) is actively developing new legislation intended to formally establish Anthropic's ethical guidelines, ensuring that human decision-making remains paramount in matters of life and death. Concurrently, Senator Elissa Slotkin (D-MI) recently introduced a bill designed to restrict the Department of Defense's capacity to utilize AI for widespread surveillance of American citizens.

Earlier this month, the Trump administration designated Anthropic as a supply-chain risk and placed it on a blacklist, following the company's imposition of limits on how the military could use its AI models. Anthropic has since filed a lawsuit, alleging that the government's action violates its constitutional rights. The company has consistently insisted that the Pentagon refrain from using its products for fully autonomous weapons and mass domestic surveillance — a position that contrasts with a deal signed by its major competitor, OpenAI. Anthropic is currently awaiting a court decision on whether to block the administration’s supply chain risk designation.

"I was alarmed to see the Pentagon take aim at Anthropic because Anthropic was simply trying to insist on policies that the vast majority of American people agree with," Schiff told The Verge in a recent phone interview. He further criticized the administration's actions, stating, "The idea that they would therefore then try to turn around and kill the company, kill one of the preeminent leaders of AI is such a hostile, dictatorial kind of an act. They would set back America’s leadership in AI, and Anthropic is one of the very best."

While Senator Schiff's office is still in the process of drafting his legislation, he has indicated its core objective is to prevent AI from being used for "certain illicit purposes." Last week, Senator Slotkin introduced a similar measure called the AI Guardrails Act, which aims to reinforce protections against domestic mass surveillance and the deployment of autonomous lethal weapons without human intervention. The precise differences or alignments between Schiff's and Slotkin's bills remain to be clarified, though both cover similar ground. Ruby Robles Perez, a spokesperson for Schiff, confirmed that his office is engaging with stakeholders and industry leaders before finalizing their bill. Slotkin's legislation specifically restricts the Department of Defense's ability to use AI to detonate a nuclear weapon or to track individuals or groups within the U.S., but it also outlines how the Defense Secretary can notify Congress in the event that "extraordinary circumstances" necessitate the use of AI to deploy autonomous lethal weapons.

In the bill Schiff is drafting, the specifics regarding what constitutes an autonomous weapon or domestic surveillance are still under discussion, but he noted that his team is also examining existing frameworks from the Biden administration. Schiff elaborated, "We haven’t resolved all of those questions yet, including how this language would apply to those who were non-citizens, but people who are lawfully in the country are deserving of protection. And then as a human rights matter, it may go beyond that as well."

He emphasized a core tenet: "We don’t want to delegate that kind of responsibility over life and death to an algorithm."

A guiding principle behind these legislative efforts is the indispensable concept of a human in the loop. Schiff asserted, "Whenever a technology has the capability of taking a human life, there needs to be a human operator in the chain of command. We don’t want to delegate that kind of responsibility over life and death to an algorithm."

Nevertheless, Schiff clarified that this does not negate the potential utility of AI in military contexts. He explained, "There are certainly circumstances in which, because AI can operate faster than human beings can, you want AI to be able to tip and cue information for human operators either that need to take steps to defend the country or that need to adjust given what it can see in real time on the battlefield." He concluded, "So the applications are very significant. They can be very beneficial from a national security and defense perspective. But they can also mean life or death. They can mean distinguishing between a civilian target and a military target, or getting those things wrong."

Given the current Democratic minority in both congressional chambers, the immediate prospects for the bill's passage may hinge on bipartisan cooperation, specifically Republicans' willingness to align with criticism of the administration. With midterm elections drawing nearer, the legislative environment is expected to become more challenging for new bills to pass before the year's end, although a shift in congressional power could occur if Democrats secure control of one or both chambers. While the proposal's unveiling is still a week or two away, Schiff is considering legislative avenues such as the National Defense Authorization Act (NDAA) to advance it.

"There’s certainly bipartisan support in the public for these kinds of limitations," Schiff affirmed. He acknowledged the political complexities, stating, "As always, you confront the issue when you’re taking steps to prevent any kind of administrative abuse that it raises issues with some of my colleagues about whether it can be read as an implicit criticism of the administration. So we’ll have to deal with that, but I’m hoping that we can make it bipartisan."

Following Anthropic's standoff with the Pentagon, OpenAI has proactively sought to justify its decision to agree to terms that have drawn public scrutiny. Despite OpenAI's subsequent assurances that it will advocate for similar restrictions, Schiff expressed a preference for legally binding mandates over reliance on institutional or corporate promises. He stated, "I would have a lot more confidence, frankly, if these were statutory requirements, than relying on the lawfulness of the Pentagon or the word of an AI CEO."