Short answer: yes, AI calling is legal in the US, with consent. In February 2024 the FCC ruled that AI-generated voices count as "artificial" voices under the Telephone Consumer Protection Act (TCPA). That means AI calls follow the same consent rules as prerecorded calls. AI calls are regulated, not banned.

This article is general information, not legal advice. Talk to counsel about your specific campaign.

What FCC 24-17 actually says

On 8 February 2024 the FCC issued a Declaratory Ruling (FCC 24-17, CG Docket 23-362) confirming that the TCPA's restrictions on "artificial or prerecorded voice" calls cover voices generated by AI, including voice clones (Cooley summary).

In practice:

  • Outbound AI calls to consumers need prior express consent, and prior express written consent for telemarketing.
  • The caller must identify itself and offer an opt-out, as with any prerecorded call.
  • Inbound calls are not affected. When a customer calls you and an AI agent answers, the TCPA's outbound rules do not apply.

Many vendor pages still say that "one-to-one" consent (written consent naming a single seller) became mandatory on 27 January 2025. It did not.

On 24 January 2025, three days before it was due to take effect, the Eleventh Circuit vacated the rule in Insurance Marketing Coalition v. FCC (Venable). The FCC later repealed it. Standard TCPA consent rules still apply.

Call typeAI allowed?What you need
Inbound: customer calls youYesNothing extra under the TCPA; disclose recording where required
Appointment reminders and service calls to existing customersYesPrior express consent
Follow-up to a lead who asked to be calledYesPrior express consent; written consent if it is telemarketing
Cold calls to consumers who never opted inNoNot permitted with an artificial voice
Calls to numbers on the Do Not Call registryOnly with an exemptionEstablished business relationship or written consent

The safest uses of AI voice are the most valuable ones anyway: answering inbound calls, following up on leads who asked, and reminding customers about appointments.

Does the AI have to say it is an AI?

At the federal level, not yet. In August 2024 the FCC proposed rules that would require AI calls to disclose themselves at the start of the call, but that is a proposal, not a final rule. Several states have their own disclosure laws, and the EU requires disclosure from 2 August 2026. See our guide to AI disclosure rules by country.

Our recommendation: disclose in the greeting everywhere. It costs one sentence and removes the risk.

AI agents usually record and transcribe every call, so recording law applies to almost every deployment.

  • One-party consent states: one participant can consent. Whether an AI agent counts as that participant is unsettled, so do not rely on it.
  • All-party consent states such as California require everyone on the call to agree. When a call crosses state lines, follow the stricter rule.

The practical answer: announce recording in the greeting ("This call is recorded and handled by an AI assistant") and let callers opt out or reach a person.

A compliance checklist for AI voice agents

  1. Use AI voice for inbound calls and for outbound calls to people who consented.
  2. Keep records of when and how each person consented.
  3. Disclose the AI and the recording in the greeting.
  4. Honor opt-outs immediately and scrub against the Do Not Call registry.
  5. Call within permitted hours in the recipient's time zone.
  6. Offer a path to a human on every call.

VoilitAI agents can disclose themselves and announce recording by default, transfer to a person on request, and log consent and call outcomes to your CRM. See Trust for how we handle data.