In plain English
- What happened
- California's B.O.T. disclosure law, in force since July 1, 2019, makes it unlawful to use an automated bot to communicate with a person online with intent to mislead them about its artificial identity when the goal is to sell goods or services or influence a vote, unless the bot's nature is disclosed clearly. Separately, the Federal Communications Commission ruled in February 2024 that AI-generated voices used in robocalls count as 'artificial' voices under the Telephone Consumer Protection Act, which already requires consent and disclosure rules for those calls. Neither is a new federal law covering every chatbot in every state.
- Why it matters to your business
- A growing number of small businesses are putting a chatbot on the website or an AI voice on the phone line to handle questions after hours, and it is easy to assume a friendly greeting is enough. If a business operates in California, or calls California numbers, the disclosure duty is not optional and applies regardless of company size. Outside those specific triggers, there is currently no single US law requiring every AI assistant to announce itself, which means the honest answer depends on where the customer is and what the bot is doing.
- Do this
- Open the chatbot or phone-agent settings and check whether the welcome message identifies it as automated, since adding one line now costs nothing and covers the strictest state rule already on the books.
The question comes up constantly as more small businesses turn on AI chat widgets and after-hours phone bots: does a customer need to be told they're not talking to a human? The short answer is that one state has already answered yes for a specific case, and federal regulators have answered yes for another. Nothing forces every business everywhere to add a disclosure banner, but the businesses most likely to get caught are the ones that assumed the question didn't apply to them.
What California's bot law actually covers
California's Business and Professions Code Section 17941, often called the B.O.T. Act, makes it unlawful to use a bot to communicate online with a person in California with intent to mislead them about the bot's artificial identity, where the purpose is to incentivize a purchase or sale in a commercial transaction, or to influence a vote. The disclosure duty only applies to that narrow situation: a bot trying to sell something or move a vote. A bot that just answers hours-of-operation questions is in murkier territory, and the law does not spell out a specific wording or placement, only that any disclosure must be clear, conspicuous and reasonably designed to inform the person.
California's bot disclosure law has applied to businesses selling online since July 1, 2019, regardless of where the company is based, if the customer is in California.
What federal rules add for phone calls
If the AI is on the phone rather than the website, a different rule applies. In February 2024, the FCC issued a declaratory ruling that AI-generated voices in robocalls are treated as artificial or prerecorded voices under the Telephone Consumer Protection Act. That means the existing TCPA consent and identification requirements, which long predate generative AI, now explicitly cover calls where an AI voice is doing the talking, not just calls using a recorded human voice.
| Date | What happened |
|---|---|
| July 1, 2019 | California's bot disclosure law (Bus. & Prof. Code §17941) takes effect for online commercial and voting-related bots |
| February 8, 2024 | FCC declaratory ruling confirms AI-generated voices in robocalls fall under existing TCPA rules |
What this means for a website chatbot or a phone tree
For a website chatbot that answers questions and occasionally suggests a product or a booking, the safest read of the California law is to treat any sales-adjacent conversation as covered, since the statute does not carve out small businesses or specify a size threshold. For an outbound AI voice call, whether it is a payment reminder or an appointment confirmation, the TCPA's consent rules apply the same way they always have, just with the added clarity that an AI voice does not get treated differently from a recorded human one.
- 1
Find the greeting text
Open the chatbot or voice-agent settings and locate the first message a customer sees or hears.
- 2
Add a plain identification line
Something like 'You're chatting with an automated assistant' at the start covers the clearest reading of California's requirement, and costs nothing to add.
- 3
Check where your customers actually are
If any meaningful share of traffic or calls comes from California, treat the disclosure as required rather than optional.
- 4
Confirm outbound call consent
If the AI places calls rather than only answering them, check that the vendor's default settings meet the consent rules that already apply under the TCPA.
- 5
Keep a copy of the vendor's compliance page
Vendor settings change; saving a dated screenshot of what was configured is more useful than remembering it later.
Reality check
A disclosure line is not a substitute for reading what the chatbot vendor actually promises about accuracy or liability if the bot gets something wrong. Adding 'you are talking to an AI assistant' satisfies the identity question; it does not cover what happens if the AI gives a customer bad information.
There is no single federal chatbot-disclosure law as of September 2026, so businesses outside California selling only to non-California customers are not bound by that specific statute. That could change, and several states have discussed similar bills, but until one passes, the honest answer for most small businesses is: check where your customers are, and if any are in California or getting a call, the existing rules already reach you.
Sources
