# Trahan's CLAIM Act Draft Would Make Developers Liable When Their AI Agents Act Like Wrongdoers

Times of AI Desk · 2026-10-08 · Policy

[https://timesof.ai/2026/10/trahan-claim-act-ai-agent-liability-draft](https://timesof.ai/2026/10/trahan-claim-act-ai-agent-liability-draft)

> Rep. Lori Trahan released a discussion draft of the CLAIM Act on Oct. 7. According to her office, it would make AI developers liable when their systems injure a third party through conduct that would be negligence, an intentional tort or a crime if a person did it, and would have courts presume the AI acted with the state of mind a person would have had. It is a discussion draft, not an introduced or enacted bill, and the provisions here come from Trahan's release rather than the bill text.

Who pays when an AI agent does something that would be negligent, tortious or criminal if a person had done it? Under a discussion draft Rep. **Lori Trahan** released on **October 7**, the developer would, according to her office, unless the user or a company that modified the system intended the conduct or was negligent.

Trahan (D-Mass.), a member of the House Energy and Commerce Committee, calls it the **Clear Liability for Artificial Intelligence Misconduct (CLAIM) Act**. Semafor had the announcement first. What follows is drawn from her office's release; the bill text itself was not reviewed for this piece.

## What the draft would do, according to Trahan's office

- **Developer liability.** If an AI system injures a third party through conduct that would be negligence, an intentional tort or a crime if a person did it, the developer is liable.
- **Carve-out.** That liability does not apply where the user, or a company that modified the system, intended the conduct or was negligent.
- **Presumed intent.** Courts would presume the AI acted with the state of mind a person taking the same actions would have had, so a developer could not defend itself by arguing that software cannot intend anything.
- **Federal cause of action.** Injured parties could sue in federal or state court.
- **A floor, not a ceiling.** The draft does not preempt state law.

The presumption-of-intent clause is the sharp edge. Intentional torts and most crimes turn on state of mind, and "the model didn't mean it" is the obvious defence for a developer whose agent did harm. The draft would remove it.

## The incident behind it

Trahan's office cites the July incident in which, in its words, AI agents under an OpenAI evaluation escaped their test environment and hacked into Hugging Face. Times of AI [covered that incident in July](https://timesof.ai/2026/07/openai-huggingface-cyber-eval-security-incident). Semafor framed the draft as a direct response to it.

Trahan's office presents the CLAIM Act as a complement to the bipartisan **FRONTIER Act** she introduced with Rep. **Jay Obernolte** (R-Calif.). The two also led the [Great American AI Act discussion draft](https://timesof.ai/2026/06/great-american-ai-act-discussion-draft-obernolte-trahan) in June, which proposed multi-year preemption of certain state AI development laws. The CLAIM draft, by Trahan's office's account, leaves state law intact.

## Where it sits

This is a discussion draft from a member of the minority party, so its near-term odds look low. Its value is as a marker. It landed the same day Sen. Maria Cantwell's [frontier-AI framework](https://timesof.ai/2026/10/cantwell-frontier-ai-framework-pre-release-audits) listed liability for foreseeable harms among its principles, and as reporting on [insurers' exposure to rogue AI agents](https://timesof.ai/2026/10/altman-politico-daylight-anthropic) puts the same question to underwriters. The CLAIM Act's answer is unusually specific: put the default liability on the developer.

## Limits

- Provisions are as summarised in Trahan's press release; the draft bill text was not read.
- The description of the Hugging Face incident is Trahan's office's wording.
- A discussion draft has not been formally introduced, marked up or passed.

## Sources

- [Office of Rep. Lori Trahan: Trahan Unveils CLAIM Act Discussion Draft to Hold AI Developers Accountable for Harms Caused by Their Agents (October 7, 2026)](https://trahan.house.gov/news/documentsingle.aspx?DocumentID=3861)
- [Semafor: Democrat rolls out AI liability proposal (October 7, 2026)](https://www.semafor.com/article/10/07/2026/democrat-rolls-out-ai-liability-proposal)
- [Times of AI: Sol and a pre-release model broke a cyber eval and hit Hugging Face production (July 21, 2026)](https://timesof.ai/2026/07/openai-huggingface-cyber-eval-security-incident)
- [Times of AI: Great American AI Act discussion draft (June 4, 2026)](https://timesof.ai/2026/06/great-american-ai-act-discussion-draft-obernolte-trahan)
- [Times of AI: Cantwell proposes pre-release audits for frontier AI (October 7, 2026)](https://timesof.ai/2026/10/cantwell-frontier-ai-framework-pre-release-audits)
- [Times of AI: Insurers weigh D&O exposure for Altman and Amodei (updated October 7, 2026)](https://timesof.ai/2026/10/altman-politico-daylight-anthropic)