Is a US Ban on AI Toys Coming? What the 2026 Federal Bills Actually Propose

Short answer: no AI toy ban is law in the United States. As of 8 September 2026, one House bill would ban AI chatbots in children's toys outright, and one Senate bill, S. 5171, has been advanced by the Commerce Committee. S. 5171 does not ban anything by itself: it orders a federal study and a joint action plan. Meanwhile the agency senators asked to step in, the Consumer Product Safety Commission, has said in writing that it lacks the authority to do so. Here is what each proposal actually says, and what has changed.

What is the AI Children's Toy Safety Act?

It is a House bill that would ban AI chatbots in children's toys. Introduced by Representative Blake Moore (R-UT), the AI Children's Toy Safety Act would prohibit the manufacture, import or sale of any children's toy incorporating an AI chatbot, 180 days after enactment. It sits with the House Energy and Commerce Committee. Moore's office cites data privacy risks and what it calls addictive and unpredictable engagement patterns as the justification.

What would S. 5171 actually do?

It would order a study, not a ban. Under the text of the Children's Artificial Intelligence Toy Safety Act of 2026 (S. 5171), the Federal Trade Commission and the Consumer Product Safety Commission would have two years from enactment to deliver a joint action plan to Congress on the safety of AI enabled toys. That plan could recommend marketing standards, safety guidelines, independent review, or temporary or permanent limits on sales, if those are needed to protect health or safety.

That distinction matters, because the bill has often been described in the press as a moratorium. As drafted, it is not one. A sales restriction is one of several outcomes the agencies could recommend at the end of a two year process, not something the bill imposes on its own. The committee's own record is at the Senate Commerce Committee, and the procedural history is tracked on GovTrack. We went through the mechanics in detail when the committee acted, in what S. 5171 would actually do.

What did the CPSC say when senators asked it to act?

It said no, and the reason it gave is the most important sentence in this whole file. In a letter dated 13 February 2026, Acting Chairman Peter A. Feldman replied to the 15 January 2026 request from Senators Klobuchar, Cantwell and Markey and set out the agency's position: the CPSC's statutory mission is built around reasonably foreseeable risks of physical injury, and it is neither equipped nor authorised to evaluate non-physical hazards such as mental, emotional or psychological harm. Extending its reach to those harms, the letter argued, would be a novel expansion beyond what Congress granted it.

Read plainly, that is a federal safety regulator saying that the specific worry parents have about a talking toy, what it says to a child and what that does to them, falls outside its remit. Physical safety of the object is covered. The conversation is not. Feldman set out the agency's broader thinking on its scope in his remarks at the 2026 Toy Fair.

This is why S. 5171 is drafted as a joint FTC and CPSC study rather than as a direct instruction. Congress appears to accept that the authority question has to be settled before anything can be enforced. It is also why, in the meantime, the practical protections available to a parent are the ones a manufacturer chooses to build in, and the ones the states are now writing into law.

How do the federal proposals compare?

Senators Cantwell, Klobuchar and Markey wrote to the Consumer Product Safety Commission in January asking it to look at AI toys under authority it already has. That route would not have needed a new law, which is why it was worth watching. The agency's February reply closed it.

Are the states moving faster than Congress?

Yes, and by a wide margin. State legislatures have been the real venue for AI toy rules in 2026, and several of those measures would bind manufacturers well before any federal bill reaches a floor vote. The Toy Association's state and federal tracker is the most complete public list.

We have covered the main ones as they landed: California's SB 867, the Oregon and Washington laws, the Maryland bill, and the wider count in 14 new state AI chatbot laws this year. If you only read one, read the state coverage rather than the federal coverage. That is where the binding rules are being written, and the CPSC's own reply explains why.

Why is this happening now?

Because advocacy groups tested products on the market this year and published what they found. Several AI toys gave answers to test prompts that no parent would accept, including responses touching on dangerous household objects. Lawmakers are reacting to specific product failures, not to the idea of a toy that talks. That is a useful distinction to hold onto, because it means how a toy is actually built matters more than whether the box says AI.

Does any of this affect Ted?

Nothing changes today, because nothing has passed. The proposals are still a reasonable checklist to hold any talking toy against. Ted has no camera, runs on topics and limits a parent sets in the Ted&Co app, and was built against EU and US toy safety certifications from the start rather than retrofitted afterwards. For the picture outside the US, we covered how the EU AI Act treats toys like Ted and what COPPA 2.0 requires. Our full security and data page sets out what is stored and for how long.

FAQ

Is it illegal to sell AI toys in the US right now?
No. No federal AI toy ban is in force as of 8 September 2026. Existing rules, including COPPA, already apply to any toy that collects personal information from children under 13.

Does S. 5171 ban AI toys for four years?
No. As drafted, S. 5171 directs the FTC and the CPSC to deliver a joint action plan to Congress within two years of enactment. That plan may recommend temporary or permanent sales limits, but the bill itself imposes none.

Can the CPSC recall an AI toy over what it says to a child?
On its own stated reading of its powers, no. Its February 2026 letter says it is neither equipped nor authorised to assess mental, emotional or psychological harm, and that doing so would require authority Congress has not given it. Physical hazards in the same toy remain squarely within its remit.

Would a ban apply retroactively to toys already sold?
As drafted, the House bill applies to future manufacture, import and sale. It does not reach toys families already own.

What should parents check before buying an AI toy today?
No camera, clear and readable data practices, working parental controls, and recognised safety certifications. Those are checkable now, whatever Congress does later. Our 2026 safety guide has the full checklist.

When will we know whether any of this becomes law?
There is no scheduled floor vote for either bill. Committee advancement, as happened with S. 5171, is a step and not a guarantee. The state laws already passed will bind manufacturers sooner than either federal proposal.

Last reviewed 8 September 2026. Legislative status changes quickly, so the linked bill texts and agency letters are the authoritative source.