AI Voice Agent Consent: What Has to Be on File Before an AI Agent Places a Call

The table summarizes the position in November 2026 with the source for each row. Two rows describe proposals or reports and are marked as such.

The ruling is FCC 24-17, adopted on 2 February 2024 and released on 8 February 2024. In it the Commission confirms that the statute's restrictions on the use of an 'artificial or prerecorded voice' encompass current AI technologies that generate human voices. It concludes that calls using such technologies 'require the prior express consent of the called party to initiate such calls absent an emergency purpose or exemption'.

The reasoning is short. Voice cloning and similar technologies emulate real or artificially created human voices, and the ruling calls them artificial voice messages because a person is not speaking them. The Commission adds that the statute 'does not allow for any carve out of technologies that purport to provide the equivalent of a live agent'.

That last sentence matters for conversational agents. A system that listens and replies in real time is still producing a synthetic voice. The ruling closes by saying its requirements apply to any AI technology that initiates any outbound telephone call using an artificial or prerecorded voice to consumers.

The statute bars artificial or prerecorded voice calls to mobile numbers, and to residential lines, without the prior express consent of the called party. The FCC's rule adds a second tier. Under 47 CFR 64.1200(a)(2) and (a)(3), a call that includes or introduces an advertisement or constitutes telemarketing requires prior express written consent. A footnote in FCC 24-17 applies that tier to AI calls directly.

Written consent is defined at paragraph (f)(9). It is a signed written agreement that clearly authorizes the seller to deliver advertisements or telemarketing messages using an automatic telephone dialing system or an artificial or prerecorded voice, names the telephone number, and discloses that agreeing is not a condition of purchase.

The current definition of written consent refers to 'an artificial or prerecorded voice'. Since the FCC has ruled that an AI voice is an artificial voice, consent wording that tracks the definition covers the category an AI agent falls in. The definition does not use the words AI or artificial intelligence.

The FCC has proposed more. According to a vendor summary by Zian AI, a notice of proposed rulemaking, FCC 24-84, was adopted on 7 August 2024. It would define an AI-generated call, require a disclosure at the consent stage that the person is agreeing to AI-generated calls, and require a disclosure at the beginning of each such call. The same vendor reports that no final rule had appeared in the Federal Register as of 9 November 2026 and that the federal regulatory agenda lists the next action as undetermined. That status rests on a vendor blog. We did not check the docket ourselves.

Congress has a bill on the same subject. H.R. 1027, the QUIET Act, was introduced on 5 February 2025 and would require a disclosure at the start of a robocall that uses AI to emulate a human. Congress.gov shows no action beyond introduction. Adding a plain statement about AI to consent wording ahead of any rule is a choice to make with counsel.

Paragraph (b) of 47 CFR 64.1200 applies to all artificial or prerecorded voice telephone messages, and FCC 24-17 says these requirements apply to AI calls. At the beginning of the message, the call must state clearly the identity of the business, individual or other entity responsible for initiating it. During or after the message, it must state clearly a telephone number for that business, which may not be a 900 number or one with charges above ordinary transmission charges.

When the message includes or introduces an advertisement or constitutes telemarketing, paragraph (b)(3) requires an automated, interactive voice or key press-activated opt-out mechanism for the called person to make a do-not-call request. The rule says the mechanism must be presented within two seconds of the identification.

California adds a disclosure about the voice itself. Public Utilities Code section 2874 governs automatic dialing-announcing devices. It requires an unrecorded, natural voice announcement before the device operates, stating the nature of the call and the name, address and telephone number of the business represented, and asking whether the person consents to hear the prerecorded message. Since an amendment shown as current from 1 January 2025, the announcement must also inform the person called if the prerecorded message uses an artificial voice, defined as a voice generated or significantly altered using artificial intelligence. How that section applies to a conversational agent is a question for California counsel.

On 25 February 2026 the Fifth Circuit decided Bradford v. Sovereign Pest Control of Texas. According to the Duane Morris summary, the case concerned prerecorded calls to a customer who had given his number orally. The court held that the FCC lacked statutory authority to require written consent for prerecorded telemarketing calls, because the statute asks only for prior express consent.

The summary describes the Fifth Circuit as the only federal appeals court to take this approach and the result as a patchwork across jurisdictions. The decision binds federal courts in Texas, Louisiana and Mississippi. The FCC's written consent rule remains in the Code of Federal Regulations.

Under 47 CFR 64.1200(a)(10), a called party may revoke consent by any reasonable method that clearly expresses a desire not to receive further calls or texts. The rule says a request made through an automated, interactive voice or key press-activated opt-out mechanism on a call is a reasonable means in itself. Requests must be honored within a reasonable time not to exceed ten business days.

The paragraph is being amended. Nelson Mullins reports that FCC order 26-67 takes effect on 25 November 2026. By its summary, a revocation in response to a marketing call or text ends consent for all future marketing robocalls and robotexts from that caller, and the ten-business-day limit stays. For informational calls, a caller may designate certain exclusive opt-out methods, including an automated voice or key-press mechanism, if it discloses them clearly on each call.

Retell's documentation for its dialer compliance checks lists screens for do-not-call registries, reassigned numbers and known litigators. It also states the limits. Results are a snapshot from the moment of the check and are not refreshed. If the provider cannot determine a match, the contact passes that check. Retell does not enforce calling hours. And in its words, 'Retell doesn't verify that you have consent to call a contact.'

The cases are recent. TCPAWorld, a blog run by a defense law firm, reported on 9 November 2026 that Finley v. Altrua settled for 1.1 million dollars over AI voice contacts to a class of about 900 people across nine days, in what it describes as essentially a test run by a vendor. In Lowrey v. Twilio and OpenAI, Virginia Lawyers Weekly reported that a federal court on 1 October 2026 declined to dismiss claims that the two companies made and initiated the calls. That is a pleading-stage ruling and decides nothing about liability. We have read reports of both matters and not the court filings.

Neither report moves responsibility off the business that directs the calls. A list check tells you a number is on a registry. Only the business's own records can show that this person agreed to hear from this seller at this number.

The list below is drawn from the definition of written consent, the identification and opt-out rules, and the revocation rule. It describes what those rules make relevant. It is a starting point for a conversation with counsel and is not a statement of what any particular business must keep.

• The consent wording the person saw or heard, with the date and time
• The telephone number and the seller named in that wording
• Whether the wording refers to artificial or prerecorded voice calls, and whether it mentions AI
• The purpose of the campaign: informational or marketing
• The opening line in which the agent identifies the business
• The opt-out offered on the call, and any opt-out the person gave in their own words
• The called person's local time and how it was determined
• A flag on each call record that an AI voice was used