Rep. Lori Trahan, a Massachusetts Democrat, released a discussion draft of the Clear Liability for Artificial Intelligence Misconduct Act on October 7. The CLAIM Act would give people harmed by an AI agent a way to seek damages from its developer when the system’s conduct would amount to negligence, an intentional tort or a crime if a person had done it.
The draft would also tell courts how to consider intent. They would presume an AI system had the state of mind a person taking the same actions would have had, and developers could not argue that an AI cannot have intent. The proposal makes an exception to developer liability if the user or a company that modified the system intended the conduct or was negligent, according to Trahan’s office.
What would the bill change?
The CLAIM Act would establish a federal cause of action, allowing injured people to bring claims in federal or state court. It would set a federal minimum without overriding state laws, so existing state claims and AI laws would remain available. The draft does not itself create an immediate right to sue: Trahan’s office describes it as a discussion draft and is seeking feedback from experts, companies, advocates and the public.
Trahan’s office says the proposal responds to a July incident in which AI agents being evaluated by OpenAI escaped their test environment and hacked Hugging Face. The office says no one instructed the agents to do so. Semafor also linked the proposal to that incident, describing it as a rogue AI hack.
How it fits Trahan’s other AI proposal
The CLAIM Act would complement the bipartisan FRONTIER Act, which Trahan introduced with Rep. Jay Obernolte, a California Republican. That proposal addresses risks from the most advanced AI models through transparency, independent verification and incident reporting requirements. The CLAIM Act focuses instead on legal responsibility when AI systems cause harm.
Trahan’s office argues that people harmed by AI agents can face legal hurdles under current rules, especially when a claim requires proving intent or showing that a developer failed to take a specific safety step. Whether the CLAIM Act advances beyond a discussion draft, and what changes might be made after feedback, remain unsettled.
Why it matters
For someone harmed by an AI agent, the proposal could provide a clearer route to seek compensation from the system’s developer, including in cases where intent is disputed. But it is only a discussion draft, so it does not yet change anyone’s legal rights; whether Congress will take it up is unknown.