That California chatbot law probably isn’t about you. Here’s how to tell.

Probably not. The California chatbot law behind the scary headlines, SB 243, governs "companion" AI built for social and emotional relationships. Its own text excludes bots used only for customer service or a business’s operational purposes. An ordinary reception, scheduling, or message-taking bot falls outside that definition on its face.

The gap between a headline and the actual statute is where owners either panic or overspend. Below is a plain-language map of the current wave of US "tell people it’s a bot" rules, who each one covers, and where a normal front desk drops out.

This is general information, not legal advice. For your business, your state, and your exact setup, talk to a lawyer.

Which states actually have an "it’s a bot" law right now?

Only a handful, and they aim at different targets. California has two (one for companion AI, one for health-care messages). Maine has a broad deception rule. Utah has two more (a general one and a mental-health one). Most states have nothing yet. Here is how they line up.

Does California’s companion-bot law (SB 243) touch my front desk?

It reaches companion AI, not service bots. SB 243 targets chatbots built to meet a user’s social needs and sustain a relationship. The statute writes the carve-out into the definition itself, so a front-desk bot is excluded twice: it isn’t a companion chatbot, and it fits the named exclusion.

From the statute (Bus. & Prof. Code Ch. 22.6, added by Chapter 677, Statutes of 2025):

"Companion chatbot" does not include any of the following: (A) A bot that is used only for customer service, a business' operational purposes, productivity and analysis related to source information, internal research, or technical assistance.

Even for a covered companion bot, the disclosure duty only fires on deception. Per §22602(a): "If a reasonable person interacting with a companion chatbot would be misled to believe that the person is interacting with a human, an operator shall issue a clear and conspicuous notification indicating that the companion chatbot is artificially generated and not human."

California also carries an older, narrower rule. The 2018 Bot Disclosure Act (SB 1001) targets bots used to incentivize a sale or influence an election vote, not front-desk help. It predates this wave and covers different ground, so if you run a sales bot in California, have counsel check that one too. This piece maps the recent laws, not every bot rule ever written.

What about a doctor’s office — does AB 3030 cover our scheduling bot?

No, if the bot handles scheduling or billing. AB 3030 reaches a health facility, clinic, physician’s office, or group practice, and only when generative AI writes patient communications about clinical information. The statute carves administrative work out by name.

From Health & Safety Code §1339.75 (Chapter 848, Statutes of 2024, effective Jan 1, 2025):

"Patient clinical information" means information relating to the health status of a patient. This information does not include administrative matters, including, but not limited to, appointment scheduling, billing, or other clerical or business matters.

There is also a human-review exemption: "If a communication is generated by generative artificial intelligence and read and reviewed by a human licensed or certified health care provider, the requirements of subdivision (a) do not apply."

Here is how those pieces stack up for a clinical message.

When the message is clinical and no licensed provider reviewed it first, the format rules also apply: for audio, spoken at the start and end; written, prominent at the beginning; continuous online or video, displayed throughout.

What does Maine require?

Maine is the broad one, but it triggers on deception, not on using a bot at all. It covers any person using an AI chatbot in trade and commerce with a consumer, with no customer-service carve-out. What matters is whether the bot poses as a person.

From 10 M.R.S. §1500-DD (PL 2025, c. 294, effective 2025):

A person may not use an artificial intelligence chatbot or any other computer technology to engage in trade and commerce with a consumer 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.

A bot that clearly presents itself as a virtual assistant, and does not pretend to be a specific person, is not the target. Maine enforces this through its Unfair Trade Practices Act.

Utah has two laws — which one is mine?

For an ordinary business, the general rule (SB 226, effective May 7, 2025) is reactive. You disclose AI when the customer asks. From Utah Code §13-75-103(1): "A supplier that uses generative artificial intelligence to interact with an individual in connection with a consumer transaction shall disclose... that the individual is interacting with generative artificial intelligence and not a human, if the individual asks or otherwise prompts the supplier about whether artificial intelligence is being used."

Proactive, up-front disclosure is required only for someone in a licensed "regulated occupation" and only for a "high-risk" use (sensitive health, financial, or biometric data, or financial, legal, medical, or mental-health advice). Utah also gives a safe harbor: an operator that clearly discloses AI status at the outset and throughout the interaction "is not subject to an enforcement action" under the disclosure section.

The second Utah law, HB 452, covers only mental-health chatbots that purport to provide therapy, and it requires disclosure "before the Utah user may access the features," again if the user has not used it within seven days, and any time the user asks. It excludes scripted tools like guided meditations and tools that only route a person to a human therapist. A general customer-service bot is not a mental-health chatbot.

So does any of this apply to you?

Run your front desk through the questions the statutes actually ask.

What does this mean when a customer calls?

At the counter, the honest move and the safe move are the same one. Have your voice agent or chat assistant say what it is in its first line: a virtual assistant for your business, not a person. That single sentence answers the trigger buried in most of these laws, which is a bot appearing human.

My stance: the cheapest compliance step is also the least clever one, and owners keep skipping it because they think sounding human converts better.

"Have your bot say it’s a virtual assistant in the first sentence, and you sidestep most of the argument before it starts."

None of that replaces a lawyer reading your exact setup and your state’s current text, and new bills keep arriving. But the front-desk version is simple. Say it’s a bot, offer a fast path to a human, and let it book the appointment. The customer gets helped, and you are not relying on a caller mistaking software for staff.