AI Regulation Tracker  /  Compliance map

The AI Chatbot Disclosure Rules Are a Patchwork. Here Is What Actually Binds Your Business

A Tennessee ban on AI posing as a therapist took effect July 1, and headlines made it sound like a wave. For a firm running an intake or support bot, the real map is narrower and stranger than the coverage suggests.

By Anthony Guerriero  ·  Reviewed by The Leveraged Years Editorial Desk  ·  Published July 6, 2026  ·  Last updated July 6, 2026

On July 1, 2026, a Tennessee law took effect, and the coverage around it suggested a new national standard for AI chatbots had arrived. It had not. The Tennessee statute is real, useful, and narrow, and the gap between what it does and what the headlines implied is exactly the kind of thing that costs a professional either needless worry or a missed obligation. Here is the map that actually governs a business chatbot.

What Tennessee actually did

Tennessee SB 1580 took effect July 1. Its rule is specific: a person who develops or deploys an artificial intelligence system "cannot advertise or represent to the public that such a system is, or is able to act as, a qualified mental health professional." A violation is a deceptive act under the Tennessee Consumer Protection Act, and the law carries a private right of action.

Note what it is not. It is not a general rule that every chatbot must announce itself as a bot. It is an impersonation ban aimed at one dangerous claim, that software is a therapist. For a therapy practice, a coaching business, or any operation whose bot brushes against mental health, it matters a great deal. For a real estate intake bot or an advisory firm's support widget, it is largely beside the point.

The laws people confuse with it

Much of the confusion comes from a second category, companion-chatbot laws, that reads like it should cover everyone and mostly does not. California's SB 243, in force since January 1, 2026, and Washington's companion-chatbot act, effective January 1, 2027, both require an operator to disclose clearly that a user is talking to an artificial system rather than a human. But these statutes target companion bots, the emotional and relationship-style products, and they carve out ordinary uses. Washington's law, for instance, excludes customer-service bots, virtual assistants, productivity tools, and games.

That exclusion is the part professionals miss. If a firm deploys a bot to book appointments, answer billing questions, or qualify leads, the high-profile companion-chatbot laws generally do not reach it. Reading SB 243 and assuming it governs your support widget leads to the wrong compliance work.

The rules that do reach an ordinary business bot

The duties that actually touch a normal intake or support bot come from general consumer-protection law, and two states lead. Maine's statute, 10 M.R.S. 1500-DD, is among the broadest in reach. It provides that a person may not use an AI chatbot to engage in trade and commerce "in a manner that may mislead or deceive a reasonable consumer into believing that the consumer is engaging with a human being unless the consumer is notified in a clear and conspicuous manner that the consumer is not engaging with a human being." That reaches any business bot capable of passing for a person.

Utah is the other. Its Artificial Intelligence Policy Act requires disclosure on a clear request, and its mental-health-chatbot law, HB 452, is more demanding still, requiring a bot to disclose that it is AI "before the chatbot features are accessed" and "any time a user asks or prompts the chatbot about whether AI is being used." Both have been in force since 2025.

The rule of thumb that survives the patchwork

Ten states, ten drafting choices, three different effective dates in this cluster alone. A professional does not need to memorize them. The through-line is simple and safe in every one of these jurisdictions. If your chatbot could reasonably be mistaken for a human, tell the user plainly and early that it is AI, and give a straight answer whenever someone asks. If your bot operates anywhere near mental health, never let it be marketed, named, or scripted as a licensed professional, because that is the specific claim Tennessee just made unlawful and that other states are moving to reach.

The wave the headlines described is not here yet. The patchwork is, and the honest version of this week's news is more useful than the dramatic one: the safest disclosure practice is already required somewhere, cheap to adopt everywhere, and the single change most likely to keep a client-facing bot out of trouble as the rest of the map fills in.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.