AI Chatbot Regulation, State by State: A 2026 Tracker

Thirty-four states are tracking some form of chatbot-specific legislation. Three have an AI companion chatbot law actually in effect. Zero have a federal one. Here's where the patchwork stands as of August 2026, and what's actually enforceable versus proposed.

Why this looks like a patchwork, because it is

Unlike data privacy, where a handful of comprehensive state laws (California, Virginia, Colorado) became the de facto national templates within a couple of years, AI companion chatbot regulation is still in its early, fragmented phase. Most of the activity across our own case coverage — the SB 243 breakdown, the Meta, Snap, OpenAI, and Google cases — sits on top of this same shifting foundation. This page is the reference version: what's law, what's litigation, and what's still a bill.

Laws actually in effect

• California — SB 243 (effective January 1, 2026). AI-identity disclosure, break reminders and content restrictions for minors, and self-harm risk protocols, enforced through a private right of action with $1,000 minimum statutory damages per violation. Full breakdown in our SB 243 article.
• New York — AI Companion Models Law (N.Y. Gen. Bus. Law § 1700 et seq., effective November 5, 2025). Requires AI-identity disclosure and self-harm detection protocols, broadly similar in structure to SB 243.
• New York — S 9051 (effective June 2026). A stricter second law that outright prohibits offering companion chatbots to minors under 18, backed by attorney-general fines of up to $25,000 per violation — materially more restrictive than the disclosure-and-safeguards model both states started with.
• Illinois — WOPR Act / HB 1806 (effective August 4, 2025). Narrower in scope than the above: it bans AI systems from independently delivering therapy or presenting as a licensed mental-health provider, rather than regulating companion chatbots generally.
• Utah — HB 452 (signed March 25, 2025). Regulates rather than bans AI mental-health chatbots specifically: requires AI disclosure, blocks the sale of user health data collected through the chatbot, and bars in-chatbot advertising.

Litigation, not legislation

Separately from the statutes above, several states have active lawsuits against specific AI companion or chatbot products — brought by state attorneys general or private plaintiffs under existing consumer-protection law, not a new chatbot-specific statute. Texas, New Mexico, and Utah have all seen suits of this kind against products covered in our case library, including Snapchat's My AI and Character.AI. This distinction matters for anyone assessing risk: a state with no chatbot-specific law can still carry real legal exposure through general consumer-protection statutes applied to chatbot conduct.

The common elements showing up across states

Despite the patchwork, the same handful of requirements keep reappearing across both enacted laws and pending bills: disclosure that the user is talking to an AI, not a human; some form of minor-specific protection, whether that's content restriction, usage limits, or an outright ban; a defined protocol for detecting and responding to self-harm risk in conversation; and, increasingly, a prohibition on a chatbot presenting itself as a licensed professional (a direct response to reporting like the case underlying our Character.AI safety review). An operator designing to this common core, rather than to any single state's exact text, is designing toward where the patchwork is converging.

The federal backdrop

No federal AI companion chatbot statute exists as of this writing. The most significant federal action to date is the FTC's Section 6(b) inquiry, opened in September 2025, seeking data from seven companies — Alphabet, Character Technologies, Meta, OpenAI, Snap, Instagram, and xAI — on their chatbots' safety practices and impact on minors. A 6(b) inquiry is an information-gathering power, not rulemaking; it can inform future FTC enforcement action or proposed legislation, but it doesn't itself create a binding national standard.

What this means if you operate one of these products

1. Treat the strictest enacted state law as your effective baseline if you operate nationally — maintaining state-by-state product variants is usually more expensive than building to the highest bar, the same pattern the industry settled into after CCPA.
2. Don't assume "no chatbot-specific law" means no exposure. General consumer-protection statutes are already being used against chatbot products in states with no dedicated statute yet.
3. Watch New York's trajectory specifically. Its move from a disclosure-based law to an outright minors ban within seven months is the clearest signal of where the stricter end of this range is heading.
4. Revisit this at least quarterly. With 98 bills active across 34 states, the map underneath this summary is one of the faster-moving areas of state law in the country right now.

Bottom line

The regulatory picture for AI companion chatbots is broad in attention and narrow in binding law — 34 states watching, 3 with something actually in effect, 0 federal. That gap is exactly where operators are making the highest-stakes, least-certain decisions right now, and it's a real factor in any build-versus-buy-versus-switch decision involving one of these products.