AI chatbot bill clears Senate panel as state fight grows

AI chatbot bill advanced in the Senate as two child-safety measures moved through committee, exposing a state-preemption fight in Congress.

The Senate Commerce Committee approved legislation that would require chatbots to delete records of interactions with minors and tell users when they are dealing with artificial intelligence. The votes give lawmakers a narrower child-safety track after broader AI legislation was pushed off until September.

Two bills leave committee

The Children’s Artificial Intelligence Toy Safety Act of 2026 would direct the National Academies of Sciences, Engineering, and Medicine to study the emotional and mental effects of children’s interactions with AI toys. The Federal Trade Commission would then have two years to send Congress a plan with legislative recommendations tied to those findings.

The committee also approved the Chatbot Act, which would create family chatbot accounts for children under 13 and require parental consent before teenagers could use chatbots. The measure would also bar targeted advertising to minors, adding a data and marketing layer to the child-safety package.

The House has already passed its own children’s online safety package, known as the Kids Act. If the full Senate takes up and passes the Commerce Committee bills, the two chambers would be positioned to reconcile their versions in a conference process.

Duty of care blocks conference

The larger obstacle is not the chatbot text alone. The House Republican majority and Democrats remain opposed to duty-of-care language in the Senate KOSA bill, which would require technology companies to mitigate foreseeable harm to users.

A senior House GOP aide, granted anonymity because the aide was not authorized to speak publicly, said the room for a House-Senate deal turns on that clause. "Whether there is a negotiation to be had will depend on whether the Senate is willing to drop the duty of care or not," the aide said.

That sentence is the clearest signal in the available record that the fight is about liability architecture, not only children’s product rules. A chatbot disclosure mandate can be implemented through interface design and data-retention systems; a duty-of-care standard can reach deeper into ranking systems, recommendation products and risk-review processes.

State chatbot laws survive

The committee also changed the preemption language in the Chatbot Act after Democrats objected. A provision that would have displaced similar state AI laws governing chatbots was removed, leaving states room to set rules that meet or exceed the federal floor.

The revised language says: "Nothing in the Act … shall be construed to prohibit … any State law, requirement, or regulation that is at least as protective of users of AI chatbots as this Act." That clause matters because it would let state lawmakers keep pressure on chatbot operators even if Congress passes a federal baseline.

The change narrows an effort by Republicans and billionaire David Sacks to secure some level of national preemption for AI chatbot rules. For large AI companies, one federal rulebook would generally be easier to operationalize than a patchwork of state requirements, especially for age-gating, parental consent and advertising controls.

Scenarios turn on Senate language

If the Senate passes the committee bills and drops or softens the duty-of-care language, a conference with the House becomes more plausible. In that scenario, the macro effect would likely be limited to compliance spending rather than broad demand; chatbot providers would face clearer federal product obligations, while the AI sector would gain a more uniform baseline for children’s services.

If the Senate keeps the duty-of-care clause intact, the House could resist a final package. That would leave chatbot companies managing existing state-level exposure and voluntary safety controls, while the wider industry would keep lobbying over whether federal AI law should preempt state rules or preserve them.

If the state savings clause remains in the final Chatbot Act, the next pressure point shifts to state legislatures and regulators. The open question is whether Congress wants a narrow child-protection statute that states can build on, or a national AI framework that limits state experiments before the market has settled around standard controls.