Most 2026 state AI-disclosure laws do not apply to AI receptionists. Washington's and California's new companion-chatbot laws both explicitly exclude bots used for customer service and business operations. California's other AI-voice law only governs outbound autodialers, not inbound calls. Utah is the exception — its AI Policy Act can reach a business phone system if a caller asks, or if the call touches financial, legal, or medical decisions.
A real wave of laws landed in 2026 — aimed at a different target
Washington signed the AI Companion Chatbot Safety Act (HB 2225) into law in 2026, effective January 1, 2027. It requires chatbots that simulate an ongoing relationship to disclose they're not human — but it explicitly excludes bots "only used for a business' operational purposes, productivity, and analysis," naming customer-service tools as the example (Fisher Phillips, 2026). California's companion-chatbot law, SB 243, does the same thing from the other coast: it requires a clear AI-identity notice when a reasonable person could be misled into thinking they're talking to a human — while excluding "chatbots used solely for customer service, internal research, or operational purposes" by name (Future of Privacy Forum, 2026).
Both laws were written for a specific problem — AI companions marketed to simulate friendship or romance, especially with minors — not for the software that answers when a customer calls a plumbing company or a dental office. Read past the headline and the carve-out is explicit in both.
| Law | Jurisdiction | Effective date | Covers AI receptionists? |
|---|---|---|---|
| AI Companion Chatbot Safety Act (HB 2225) | Washington | Jan 1, 2027 | No — names customer service as excluded |
| Companion Chatbot Law (SB 243) | California | In force | No — excludes customer service/operational bots |
| Artificial-voice disclosure (Pub. Util. Code §2874 / AB 2905) | California | Jan 1, 2025 | No — outbound autodialers only |
| AI Policy Act (SB 149, amended by SB 226) | Utah | May 7, 2025 (amendment) | Yes — on request, or in high-risk interactions |
The one California law that's actually about phone calls — and still isn't about your receptionist
California Public Utilities Code §2874 is the closest thing to a real "AI phone disclosure law," and it does require callers to state if a message uses an artificial voice. But the opening line of the statute gives away its scope: "Whenever telephone calls are placed through the use of an automatic dialing-announcing device..." (FindLaw, 2026). That's an outbound autodialer blasting a prerecorded message — a robocall — not an AI system answering an inbound call from a customer who dialed you. The distinction matters because several vendor blog posts covering this law blur outbound and inbound together; the statute itself doesn't.
Utah is the one that actually reaches a receptionist
Utah's AI Policy Act (SB 149) is the outlier worth knowing. As amended by SB 226, it requires a business to disclose that a consumer is interacting with AI in two situations: when the consumer or supplier directly asks, or when the interaction is "high-risk" — meaning it touches financial, legal, medical, or mental-health decisions, or sensitive personal data (FPF Utah, 2026). A phone system booking appointments or answering pricing questions is a lower-risk case than most; the safe harbor the law provides is straightforward — disclose clearly at the start of or during the interaction, and you're covered.
Federal rules? Still a proposal, not a rule
The one federal rule already in force is narrower than most business owners assume. The FCC's Declaratory Ruling (FCC 24-17), effective February 8, 2024, put AI-generated voices under the TCPA's existing rules for artificial and prerecorded voice calls — meaning prior express consent is required before placing that kind of call (Resemble AI, 2026). That ruling is about outbound calling, the same category as California's robocall law. A separate FCC proposal from mid-2024 that would add real-time AI-disclosure requirements is still an open rulemaking, with no final rule adopted as of this writing (Resemble AI, 2026).
Montana has no AI phone-disclosure law at all
Montana's AI-specific consumer law is the Consumer Data Privacy Act's AI provisions, effective October 1, 2024 — and it's aimed at a different problem entirely: it requires disclosure when AI makes a "consequential decision" affecting a consumer's access to credit, pricing, or services, plus an opt-out to request human review (ai-law-tracker.com, 2026). It doesn't reach a general conversational phone system, and Montana has no separate law requiring a business phone AI to announce itself. For a Kalispell contractor or a Missoula dental office, there's currently no state disclosure mandate either way.
- If a caller directly asks whether they're talking to a person, say so plainly — this satisfies Utah's law and is the safe move everywhere else, law or no law.
- Don't script the system to actively claim it's human if asked — that's the behavior every one of these laws, current or proposed, is actually trying to stop.
- If your business books calls from customers in multiple states, Utah's request-based trigger is the strictest rule likely to touch you — build the disclosure habit once and it covers every jurisdiction, not just Utah.