Search "does an AI receptionist have to disclose it's AI" and you'll find dozens of vendor blog posts citing state laws, complete with bill numbers and effective dates. We pulled the actual bill text and law-firm analysis behind those claims to write this piece — and several of the most-cited "laws" don't say what the blogs claim, or don't exist at all. Here's what's actually on the books as of July 2026, and what genuinely applies to a service business running an AI receptionist.
The Federal Picture: FCC's Ruling Covers Outbound Calls, Not Inbound Ones
On February 8, 2024, the FCC issued a declaratory ruling confirming that an AI-generated voice counts as an "artificial voice" under the Telephone Consumer Protection Act (TCPA). That's real, and it matters — but it applies to calls a business places to consumers using a synthetic voice, like AI-voice robocalls or outbound follow-up campaigns. It does not govern a system that simply answers when a customer calls your business.
There is currently no comprehensive federal law requiring a business to announce "you're talking to an AI" when a customer calls in and the AI receptionist answers. If your AI system only handles inbound calls, the FCC ruling isn't the thing to worry about. If it also places outbound AI-voice calls — appointment reminders, missed-call follow-up in a synthesized voice — that's where standard TCPA consent rules apply, the same as any other automated call.
State Laws: A Real Patchwork, Narrower Than the Blogs Claim
We checked bill text and law-firm coverage for the states most often cited. Here's what's actually true for each, as of July 2026:
| State / law | What it actually requires | Applies to a typical AI receptionist? |
|---|---|---|
| Utah — SB 226 (effective May 7, 2025) | Must disclose it's a bot only if the consumer directly asks, or during a "high-risk" interaction (financial, legal, medical, or mental-health decisions involving sensitive personal data). | Conditionally — mainly if you handle sensitive personal/financial data over the call. |
| Texas — TRAIGA / HB 149 (effective Jan 1, 2026) | Disclosure obligations apply to state agencies and to healthcare providers using AI in diagnosis or treatment. Private employers/businesses are not required to disclose AI use to customers or employees. | No, for most home-service and trade businesses. |
| California — SB 1001 (B.O.T. Act) (effective 2019) | Requires bot disclosure only on public-facing websites/apps with 10M+ monthly U.S. visitors, for sales or election-influence communication. Doesn't cover phone calls or small-business systems. | No — scope is limited to large online platforms, not phone systems. |
| Nevada | Nevada's 2025 AI laws cover AI in mental/behavioral healthcare (AB 406) and AI-generated content in paid political ads (AB 73). We found no signed Nevada law requiring general customer-service AI disclosure — despite several SEO sites citing a "Nevada SB 149 disclosure law" that doesn't match any bill we could verify. | No — and be skeptical of sites claiming otherwise; we couldn't verify that bill. |
Colorado's AI Act, which does regulate "high-risk" AI systems more broadly, is currently set to take effect June 30, 2026 — worth watching if you operate there, though its consumer-disclosure provisions are aimed at consequential decisions like credit, employment, and housing, not routine call answering.
Why the Confusion? Vendor Blogs Are Citing Each Other, Not the Law
Most of the "AI voice disclosure law" roundups ranking for this topic are written by AI-receptionist vendors, and they cite each other rather than primary sources. In our research, one widely repeated claim attributed a voice-disclosure requirement to a California bill (AB 1018) that is actually about automated decision systems in employment and healthcare — and was placed on the inactive file in September 2025, meaning it isn't even law. If you're making a compliance decision based on a blog post, check the bill number against your state legislature's own site first.
Even Where It's Not Required, Disclosure Is the Safer Default
Legal requirement aside, consumer sentiment is worth factoring in. Recent surveys on AI customer service consistently find a strong majority of people say they'd prefer to speak with a human, and that frustration with AI agents is rising. But the research also points to what actually damages trust: being misled, not simply talking to an AI. A short, plain disclosure at the top of the call — something like "Thanks for calling [business] — I'm the AI assistant, I can get you booked or grab your info for the team" — tends to cost less trust than a customer figuring out later they were talking to a bot that didn't say so.
It's also the direction state law is clearly heading: three different states now have some form of AI disclosure requirement on the books, even if each is narrower than the vendor blogs suggest. Building the habit now costs nothing and removes the question entirely as more states catch up.
The Bottom Line
As of July 2026, no state or federal law requires most service businesses to announce "I'm an AI" on every inbound call. Utah is the one state with a real, if narrow, disclosure trigger. Everything else circulating about California, Texas, or Nevada "AI phone disclosure laws" either applies to a different context entirely (healthcare, elections, large platforms) or doesn't check out against the actual bill text. That said: this is not legal advice, state AI legislation is moving quickly, and the accurate answer for your specific state can change with the next legislative session — confirm current requirements with an attorney before treating any blog post, including this one, as the final word.