AI-Voice Calls and the TCPA: What Businesses Need to Know

An AI-generated voice is an artificial voice under the Telephone Consumer Protection Act, according to the FCC, so an AI sales or service call to a cell phone needs the called person’s prior express consent, the same as a recording. Ringless voicemails count as calls too. California adds its own disclosure rule for recorded-message dialers that use AI voices. For businesses deploying voice AI, consent records and vendor contracts are now the core of TCPA risk.

Why the FCC treats AI voices as recordings

The TCPA bars calls to a cell phone using “an artificial or prerecorded voice” without prior express consent, under 47 U.S.C. § 227(b)(1)(A)(iii). In February 2024, the FCC confirmed in FCC 24-17 that those restrictions “encompass current AI technologies that generate human voices.” The ruling states the consequence directly: “Therefore, callers must obtain prior express consent from the called party before making a call that utilizes artificial or prerecorded voice simulated or generated through AI technology.”

The FCC also closed the obvious workaround. A voice agent that converses in real time can feel like a live caller, but the ruling says it “also makes clear that the TCPA does not allow for any carve out of technologies that purport to provide the equivalent of a live agent.”

After the Supreme Court’s 2025 decision in McLaughlin Chiropractic Associates v. McKesson Corp., courts aren’t bound by FCC rulings in private cases, so a defendant can argue the ruling went too far. Our post on McLaughlin v. McKesson covers that decision. Don’t build a calling program on that argument. The statute’s word is “artificial,” and a machine-generated voice is a hard thing to call anything else. Consent is the safer footing.

Ringless voicemail is a call

Ringless voicemail drops a recorded message into a voicemail box without ringing the phone. The FCC rejected the argument that it isn’t a call, finding in FCC 22-85 “that “ringless voicemail” to wireless phones requires consumer consent because it is a “call” made using an artificial or prerecorded voice.” The FCC also denied the request for a retroactive waiver. Our post on defending ringless voicemail claims covers the arguments that remain.

Collection agencies face a second rule. The CFPB’s official interpretation of its debt collection rule says “placing a telephone call” includes conveying a ringless voicemail, so those messages count toward Regulation F’s call-frequency limits, per the CFPB’s commentary.

Platform and vendor liability

Voice-AI calls often involve several companies: the business selling something, the dialer, and the company supplying the voice model. A Virginia federal court decided in October 2026 that claims could proceed against both the number provider and the voice provider. In Lowrey v. Twilio, the court summarized the complaint this way: “As Lowrey alleges, Twilio’s and OpenAI’s platforms work together to place the calls and to provide the artificial voice that engages with the consumer.” It concluded that “there is also an adequate statutory basis to hold OpenAI liable for lending its technology to telemarketing calls.”

That ruling came at the pleading stage, accepting the complaint’s allegations as true, and it didn’t find that either company violated the law. The same order dismissed the Do Not Call claims, holding that texts aren’t telephone calls and that the Do Not Call rules reach only residential landlines. Those holdings bind no court in California, but they’re arguments a California defendant can make.

Vendor allocation of TCPA risk is about to become a standard negotiating point. Businesses buying voice AI should ask who captures consent and who indemnifies whom. Platforms should ask the same questions of their customers. Our post on TCPA terms in vendor contracts covers the clauses to ask for.

California’s AI-voice disclosure rule

California regulates automatic dialing-announcing devices, meaning dialers that play recorded messages. Under Public Utilities Code § 2874, “the device may be operated only after an unrecorded, natural voice announcement has been made to the person called by the person calling.” Since January 1, 2025, that live announcement must also “Inform the person called if the prerecorded message uses an artificial voice.” The statute defines an artificial voice as “a voice that is generated or significantly altered using artificial intelligence.”

California’s Consumers Legal Remedies Act also lists as unlawful “Disseminating an unsolicited prerecorded message by telephone without an unrecorded, natural voice first informing the person answering the telephone of the name of the caller or the organization being represented,” under Civil Code § 1770(a)(22). It exempts calls to people with an established relationship, collection calls, and calls the recipient requested.

Federal identification rules for recorded messages

Every artificial or prerecorded message must “At the beginning of the message, state clearly the identity of the business, individual, or other entity that is responsible for initiating the call,” under 47 C.F.R. § 64.1200(b)(1). The Lowrey court held that this rule doesn’t support a private lawsuit on its own. It’s still the rule the FCC enforces, and a voice agent that won’t say whose product it’s selling invites scrutiny.

Compliance and exposure at a glance

Before a voice-AI claim arrives, or right after

1. Preserve call detail records from the dialer and the voice platform, the audio or transcripts for each call, and the consent record tied to each number, under a litigation hold once a claim arrives.
2. Pull the opt-out handling: every key-press or spoken opt-out, how the AI agent responded, and when the number was suppressed.
3. Pull Registry scrub records for any sales calls and the version dates used.
4. Collect the contracts with the voice-AI provider, the dialer, and any lead sources, including consent-capture and indemnity terms.
5. Talk to a lawyer before responding to a demand letter, contacting the caller’s lawyer, or changing the platform’s settings in ways that could look like an admission.

For claims involving AI-voice calls and ringless voicemail campaigns, the firm’s California TCPA defense page sets out what the defense looks at first.

AI calling questions from compliance teams

Does consent for live calls cover AI-voice calls?

Not necessarily. Prior express consent has to cover calls using an artificial or prerecorded voice. Consent language written for live agents may not reach them, and for telemarketing the FCC’s rules call for prior express written consent. Review your consent forms for language that expressly covers automated and artificial-voice calls before switching a campaign to voice AI.

Is our voice-AI vendor liable instead of us?

Possibly both of you. The seller is usually the first target, and Lowrey shows plaintiffs will also name the platforms. Contracts decide who bears the cost between the parties, but they don’t stop a plaintiff from suing either one. Allocation of consent responsibilities and indemnity should be settled before launch. Our post on texting platform liability covers how courts sorted platforms and their customers before Lowrey.

Can an AI agent handle opt-outs?

It can, if it’s configured to recognize and honor them. The FCC’s rules treat an automated, key-press or voice-activated opt-out on a call as a reasonable way to revoke consent, and companies must honor revocations within ten business days. An agent that talks past a customer’s request to stop is a lawsuit waiting to happen. Our post on AI text chatbots and the TCPA covers the same problem when the conversation happens by text.

Do California’s disclosure rules apply to calls from out of state?

California’s dialer rules apply to devices operated to place calls received in California. A company outside the state that calls California numbers with recorded or AI messages should plan for the live-introduction and AI-disclosure requirements, along with the federal consent rules that apply everywhere.

Defending claims over AI-voice calls

If your company uses AI voice or ringless voicemail platforms and has been named in a TCPA claim, how and when consent was captured is usually the first question. Rosenberger + Kawabata represents businesses defending TCPA lawsuits and class actions through its business litigation practice, including claims involving AI-voice calls and ringless voicemail campaigns. Contact Rosenberger + Kawabata online for a confidential consultation, or call (310) 894-6921.

Sources

The federal statute is 47 U.S.C. § 227, and the FCC’s rules are at 47 C.F.R. § 64.1200, as published in the eCFR through October 7, 2026. The FCC rulings are FCC 24-17 on AI voices (Feb. 8, 2024) and FCC 22-85 on ringless voicemail (Nov. 21, 2022). The CFPB’s guidance is its official interpretation of 12 C.F.R. § 1006.14. California’s rules are Public Utilities Code § 2874 and Civil Code § 1770. The court decisions are McLaughlin Chiropractic Associates v. McKesson Corp., 606 U.S. 146 (2025), and Lowrey v. Twilio, Inc., No. 6:25-cv-00116 (W.D. Va. Oct. 1, 2026).

This post is general information about the Telephone Consumer Protection Act and related laws as of October 8, 2026. It is not legal advice, and reading it does not create an attorney-client relationship. Courts and the FCC change the rules in this area often, and the outcome of any claim depends on its facts. Past results do not guarantee future outcomes.