Senators Josh Hawley (R-MO) and Chris Murphy (D-CT) are preparing bipartisan legislation to address liability for artificial intelligence systems. The question at the heart of the effort is easy to pose and difficult to resolve: when an AI system causes harm, who is responsible for paying? Today there is no clear federal framework that answers it, and closing that gap is what this legislative push is meant to do.

The proposals already on the table

The planned Hawley-Murphy bill would arrive amid a series of AI accountability measures that Hawley has already helped push through the introduction stage. Two stand out.

The first, the AI LEAD Act (S. 2937), was introduced on September 29, 2025, with Sen. Dick Durbin (D-IL) as co-sponsor. The measure would classify AI systems as products under federal law, allowing victims to file product liability claims against companies whose systems cause harm. The classification is the crux: product liability has traditionally applied to tangible goods, leaving software developers room to argue their systems fall outside it. Its stated objective is to push developers toward safer design and deployment by ensuring that negligent engineering carries real financial consequences. The bill has drawn endorsements from the American Association for Justice and the National Center on Sexual Exploitation.

The second, the GUARD Act, was introduced on October 28, 2025. It would ban AI companions for minors outright and require chatbots to disclose that they are not human. The legislation also sets out strict new rules, including criminal penalties for companies that produce harmful content aimed at children. RAINN, the anti-sexual violence organization, has backed the GUARD Act, as have parents affected by their children's interactions with AI.

The legislative backdrop

Congress has held hearings in which affected individuals described harms tied to AI systems, and those testimonies have shaped the legislative push. The discussions have touched on some of the biggest names in the field, including OpenAI and Google.

Hawley has also made the case publicly. In an op-ed, he argued that companies should be held accountable for reckless AI design and called for a rethink of how technology firms approach building these systems.

As of early 2026, neither the AI LEAD Act nor the GUARD Act has become law; both remain under legislative consideration. The road from introduction to statute is a long one: a bill must survive committee review, win votes in both the House and the Senate, and be signed before it takes effect — a process where cross-party sponsorship is typically an asset rather than an obstacle.

What a liability framework would change

For AI developers, a product liability framework would turn safety from a public relations talking point into a line item on the legal budget — and, if the AI LEAD Act's stated aims carry over, into decisions about how systems are designed, tested, and documented before release. Bills such as the AI LEAD Act aim to close the classification gap by settling the question in federal law, removing companies' ability to argue that their systems are services, platforms, or tools rather than products.

The details of the Hawley-Murphy bill will determine whether it overlaps with the AI LEAD Act, builds on the GUARD Act, or carves out new ground. Key questions include how it defines harm, who can sue, and whether it includes criminal penalties or sticks to civil liability. None of those answers are public yet; the earliest signals will come from the bill's actual text, its committee assignment, and whether Congress schedules hearings on it.