AI Phone Agents and Call Recording Consent Laws

An AI that answers your phone usually records and transcribes the call. Two sets of rules apply: recording consent, which varies by state, and AI disclosure, which only a few states have written into law. Here is what the primary sources say, as of October 2026.

General information, not legal advice. This article describes what statutes, court decisions and government documents say, as of October 2026. It does not tell you whether your phone setup complies, and these laws are changing. Confirm your greeting and recording practices with counsel, especially if you take calls from more than one state.

An AI receptionist is a recording device with a voice. Most of these systems record the call, transcribe it and store it so your team can see what the caller wanted. That puts you under call-recording law whether or not you ever think of it that way. Separately, a handful of states now have written rules about telling people when they are talking to AI.

This guide covers three questions: whose consent a recording needs, what the law says about disclosing an AI voice, and what a call greeting can contain. It sticks to sources we read, and it says where we could not find one. The buying side of AI answering is in our buyer's guide to AI receptionists for contractors, and the texting rules sit in the companion TCPA guide.

The federal floor

The federal wiretap statute makes it lawful for a person who is a party to a call to record it, or to record it where one party has consented, unless the recording is made to commit a criminal or tortious act. The text is in 18 U.S.C. 2511(2)(d). The Reporters Committee for Freedom of the Press describes this as the "one-party consent" minimum that applies across the country, and notes that states may impose stricter rules.

Many do. That is why a business that records calls usually cannot rely on the federal rule alone.

State laws vary, and the better sources keep changing as legislatures and courts act, so we are linking a maintained list instead of building our own. The Reporters Committee's Reporter's Recording Guide has a page for every state, and the introduction groups them as follows (read October 2026):

The guide is written for journalists and says it should not replace advice from a lawyer. Read the state page you care about and ask counsel before relying on it. We did not find an authoritative source on which state's law governs a call between people in two states with different rules, so we are not giving a rule of thumb for that.

California

California is the all-party state contractors most often ask about, and it has two relevant sections:

According to the Reporters Committee's summary, a first offense under these sections is punishable by a fine of up to $2,500 or up to one year in jail, and a California appellate court has held that even recording only one side of a call could violate the law if consent was not obtained from all participants. We read the Reporters Committee's California page rather than the code section itself, because the legislature's site blocked automated access, so check the code text with counsel.

Florida

Florida also requires everyone's consent. Section 934.03(2)(d) of the Florida Statutes makes interception lawful when "all of the parties to the communication have given prior consent." Section 934.03(4)(a) classifies a violation of the interception prohibition as a third-degree felony, with a lesser classification for narrow circumstances listed in (4)(b). The Reporters Committee's Florida page also describes civil remedies under section 934.10 and notes that the law covers text and email messages between devices.

Most AI answering greetings do not ask "do you consent?" They say the call is recorded and continue. The Reporters Committee's introduction states that it is generally legal to record any type of conversation when the parties are warned of the recording and continue with the conversation, because consent is presumed in those instances (it cites a federal appeals court decision, Alexander v. Pathfinder, Inc., 8th Cir. 1999), and adds that it is a best practice to record the subject's verbal consent.

Two cautions from that same source: it says "generally," and it is describing a general principle across states, not a ruling about any single state's statute. Ask counsel whether notice plus continuing is enough in each state where your callers live, or whether you want an affirmative "yes."

Part 2: AI voice disclosure

Federal: the FCC's 2024 ruling

On February 8, 2024, the FCC announced the unanimous adoption of a Declaratory Ruling that recognizes calls made with AI-generated voices as "artificial" under the Telephone Consumer Protection Act. The FCC's announcement says the ruling took effect immediately and that it holds those calls to the same standards as other artificial or prerecorded voice calls, including the requirement for prior express written consent before telemarketing robocalls.

Note what that ruling is about. The FCC announcement frames it around robocalls, meaning calls a business places using an artificial or prerecorded voice. An AI that answers a call a customer places to you is a different fact pattern, and we did not find an FCC or federal statement that addresses it. If your AI places outbound calls, such as a callback to a web lead, the consent rules in our TCPA guide are the ones to read first. Also note the FCC's announcement is a press release; the ruling's full text is the official source.

We found no federal statute or rule requiring a business to announce that an inbound caller is speaking to AI. We did not search every agency, so treat that as "not found," not "does not exist."

State laws we could confirm are enacted

The state picture is thin, scattered and moving. We list only laws we read in enacted form. Bills do not count.

Utah (enacted). S.B. 226 (2025), effective May 7, 2025, enacted section 13-75-103 of the Utah Code. It provides that a supplier that uses generative AI to interact with an individual in connection with a consumer transaction must disclose that the individual is interacting with AI and not a human if the individual asks or otherwise prompts the supplier about whether AI is being used. The question must be a clear and unambiguous request to determine whether the interaction is with a human or AI. People in regulated occupations, which means licensed or certified by Utah's Department of Commerce, have a more demanding duty in high-risk interactions (health, financial, biometric data and similar), including disclosing verbally at the start of a verbal interaction.

The same law, in section 13-75-104, provides a safe harbor from enforcement if the AI clearly and conspicuously discloses at the outset of the interaction, and throughout, that it is generative AI, is not human, or is an AI assistant. Penalties include an administrative fine of up to $2,500 per violation. The bill defines "supplier" by reference to a Utah consumer statute we did not read, so whether it reaches an out-of-state contractor taking a Utah caller is a question for counsel.

Maine (enacted). L.D. 1727, approved June 12, 2025, adds 10 M.R.S. section 1500-Y. It says a person may not use an AI 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 they are not. The law defines "AI chatbot" as a program that simulates human conversation through textual or aural communications. A violation is a violation of the Maine Unfair Trade Practices Act. We did not verify the effective date from the bill text.

California (enacted, narrower). A California bill, AB 2905, was chaptered September 20, 2024 as Chapter 316. According to the summary on that page, it requires the natural-voice announcement that precedes certain prerecorded messages to also inform the person called if the message uses an artificial voice. We saw a summary, not the enacted code section, and it concerns announcements before prerecorded messages, so it does not obviously cover a live two-way AI conversation. We are including it so you know to ask about it, not to say how it applies.

Other states have chatbot and AI bills and some have statutes we did not read. We left them out rather than guess. If you answer calls in a state not listed here, ask counsel to check it.

Part 3: What a call greeting can contain

The sources above point to the same short list of elements. Whether any of them is enough for you is a call for counsel. This is wording to bring to that conversation, and it is illustrative only, not reviewed by a lawyer.

Illustrative greeting: "Thanks for calling Example Heating and Air. This is an AI assistant, and this call is recorded so we can take down your request. If you would rather talk to a person, say 'representative' at any time."

What each part is for, with its source:

Two design points that follow from the text:

  1. Put the recording notice at the start of the call. The Reporters Committee's point about presumed consent rests on people being warned and then continuing, which only works if the warning comes first.
  2. Check that the AI never claims to be human. Maine's statute is aimed at conduct that may mislead a reasonable consumer into thinking they are talking to a human, and Utah's duty is triggered by a direct question.

If you transfer to a human, tell the caller, and apply the same recording notice to the human side of the call.

For how an AI front desk compares with a human answering service on these points, see AI receptionist vs answering service for contractors. Greeting and call-flow design is one piece of AI customer support implementation, and your lawyer should review the final wording.

Frequently asked questions

Do I need to tell callers an AI is answering?

No federal statute or rule we found requires it for inbound calls. Utah requires disclosure when a caller clearly asks whether they are talking to AI, with a safe harbor for upfront disclosure. Maine bars using AI to mislead a reasonable consumer into thinking they are dealing with a human unless the consumer is clearly notified. Other states may have laws we did not verify.

Both states require the consent of all parties to record a phone call under the statutes and summaries cited above: California Penal Code sections 632 and 632.7, and Florida Statutes section 934.03. Penalties are criminal and civil. Ask counsel what notice or consent your system should collect.

Did the FCC say AI voices are illegal on calls?

The FCC's February 2024 ruling says AI-generated voices are "artificial" under the TCPA, so robocalls using them need the consent the TCPA requires. The FCC's own announcement describes the ruling in those terms. It is not a general ban on AI voices.

Where can I check my state's recording law?

Use the Reporters Committee's Reporter's Recording Guide for a page on each state, then read the statute it cites or ask your attorney. We link it instead of listing every state because the underlying laws change.

Want us to look at your numbers?

Tell us where leads are slipping. A founder reads every message and replies personally. The first call is free.

Thanks, it's in.

We'll be in touch soon. Prefer email? dihydigital@gmail.com

Ready to find the revenue leak?

Book a free 30-minute strategy call. We'll look at where leads are slowing down, where follow-up is leaking, and what it would take to recover more booked work.

Book a Free Strategy Call →