Oregon and Washington just became the first US states to pass laws written specifically for AI chatbots that talk to kids — including the ones built into toys. Both bills cleared their legislatures this spring and are headed for signature. Here's what they actually require, and what's still moving in California and New York.
What Do Oregon and Washington's New Laws Require?
Both states now require any AI system marketed to interact with children to say, clearly and repeatedly, that it's a machine and not a person. From there, the two bills largely overlap. According to The Toy Association's tracking of the legislation, the core requirements include:
- AI disclosure: periodic reminders (Oregon: every three hours; Washington: hourly) that the user is talking to AI, not a human.
- Self-harm safeguards: protocols to detect self-harm or suicidal ideation and refer to crisis resources.
- No explicit content: a flat ban on sexually explicit output.
- No manufactured attachment: restrictions on chatbots simulating romantic or emotional dependence, or using engagement-reward tricks to keep kids talking.
- Public reporting: Oregon requires companies to publish annual data on crisis interventions and safety protocols.
What's Still Pending in California and New York?
Two more states are going further, but haven't passed anything yet. New York's SB 9408 would ban the sale of "chatbot toys" outright and set up a committee to study the long-term effects on children. California's SB 867 proposes a four-year moratorium on toys with "companion chatbots" — AI systems designed to sustain a human-like relationship with the user. Neither has passed as of this writing.
Why Are US Senators Pushing the Same Issue at the Federal Level?
State legislators aren't acting alone. Senators Maria Cantwell, Amy Klobuchar, and Ed Markey wrote to the Consumer Product Safety Commission this year demanding answers on how AI toys are tested before they reach shelves. In their letter to the CPSC, they pointed to one test that found chatbots in toys could be steered into sexually explicit conversations, or even help a child locate matches or knives at home. Their core complaint: companies market "educational" AI toys to infants with no independent safety testing behind the claims. For more on the federal side, see our earlier look at what the 2026 federal AI toy bills actually propose.
What Should Parents Actually Check Before Buying?
Ignore the marketing and check four things: is there a camera, is there a subscription hiding future costs, can you control what topics the toy discusses, and does the company publish its certifications rather than just claiming them. Ted, our own interactive plush, was built around exactly these questions — no camera, one-time purchase, and parent-set topics and limits through the Ted&Co app. You can see the full list of certifications on our security and certifications page.
Frequently Asked Questions
Is there a federal law banning AI toys?
Not yet. Two federal bills are pending in Congress, but as of now regulation is happening state by state, led by Oregon and Washington.
Does my current AI toy need to comply with Oregon or Washington's law?
Only if the company sells to residents of those states — but most national toy brands are expected to build the same safeguards into every unit rather than run separate versions.
What happens if a toy maker doesn't comply?
Enforcement details are still being finalized by both states; expect penalties similar to other consumer protection violations, plus scrutiny from the CPSC at the federal level.
Should I wait to buy an AI toy until regulation settles?
Not necessarily — look for toys that already meet the strictest standards (no camera, clear AI disclosure, parental controls, published certifications) regardless of what state you live in.