Are AI Phone Calls Legal? Federal Rules and Recipient Rights

AI phone calls are not illegal by themselves, but since February 8, 2024 the Federal Communications Commission (FCC) treats an AI voice as an "artificial" voice under federal robocall law.

At Layer3 Labs, we build customer-service automation for small businesses, including the phone lines that now take calls from customers' AI agents.

Under current federal standards, placing an AI call to a consumer cell phone or residential line requires prior consent, and noncompliant callers face statutory damages ranging from $500 to $1,500 per violation.

Legal Status of AI Phone Calls Under Federal Law

An AI phone call is not illegal by itself, but since the FCC's February 2024 ruling an AI voice counts as an "artificial" voice under federal robocall law. That means AI calls to cell phones or homes need prior consent (written consent for sales calls), and you can sue over unwanted ones for $500 per call, or up to $1,500 if the caller acted willfully. A customer's own AI agent calling your business about that customer's booking is not a robocall campaign. We have not verified any court or FCC ruling on that use.

One primary statute governs synthetic voice technology. That is the Telephone Consumer Protection Act (TCPA, 47 U.S.C. § 227).

This legal distinction matters because recipients possess concrete private rights of action against entities placing unconsented calls. Organizations managing outbound campaigns need dedicated compliance architectures. We outline those architectures in our compliance guide for automated outbound dialing.

The FCC Declaratory Ruling on Synthetic Voices

On February 8, 2024, the Federal Communications Commission (FCC) unanimously adopted FCC Declaratory Ruling FCC 24-17, resolving ambiguity regarding synthetic speech systems. The agency clarified that technologies generating humanlike voices fall squarely within existing statutory restrictions.

The ruling establishes that AI-generated voices are artificial voices under the TCPA. Callers cannot circumvent federal robocall restrictions simply by deploying dynamic, real-time natural language models instead of static audio recordings. The physical generation mechanism does not alter the regulatory definition. The full ruling text confirms that conclusion.

Because of the ruling, the TCPA's consent rules and its right to sue apply to AI-voice calls. They apply the same way they do to prerecorded robocalls. Any outbound telemarketing system that produces synthetic speech must clear identical regulatory thresholds to traditional prerecorded phone systems.

Consent Standards for Different AI Call Categories

The legality of any AI-driven call depends on the line type reached and the underlying purpose of the outreach. Federal regulations impose tiered consent thresholds between commercial marketing campaigns and operational transactional notifications.

Reaching a personal cell phone using an artificial voice requires prior express consent for informational notices. Commercial sales pitches demand formal written documentation before dialing.

Business landlines occupy a distinct federal legal category. The TCPA artificial voice consent restrictions were drafted around residential lines and wireless phones. Federal robocall protections largely exclude commercial landline numbers.

• AI voice calls to personal cell phones require prior express consent, except when dialed for bona fide emergency purposes.
• Telemarketing calls using AI voices require prior express written consent before initiating contact.
• AI voice calls placed to residential landlines require prior express consent unless a specific regulatory exemption applies.
• Calls to business landlines are largely outside the TCPA's artificial-voice consent rules, so ask a lawyer about your specific line.
• A customer's own AI agent calling a business about that customer's booking is not a telemarketing campaign, but no court or FCC ruling on that use has been verified.

Recipient Rights and Private Rights of Action

Recipients of unlawful synthetic voice calls hold actionable enforcement rights under federal statute. If an automated system dials your wireless phone without consent, you can tell the caller you revoke any consent you gave and ask to be put on its do-not-call list.

Consumers can register personal numbers on the National Do Not Call Registry to curb commercial solicitations. The Federal Trade Commission (FTC) oversees the registry. The agency notes that business-to-business telephone numbers are generally excluded from registry enforcement.

Regulatory enforcement agencies accept reports of unlawful calls. Victims of deceptive or abusive calling campaigns can lodge complaints through the FTC fraud reporting portal. They can also file formal notices with the FCC consumer complaint center.

Autonomous Customer Agents Calling Local Businesses

A distinct operational category has emerged through consumer-facing AI assistants that place phone calls to local merchants on behalf of real individuals. These systems handle localized administrative tasks, such as confirming inventory availability, booking salon appointments, or reserving restaurant tables.

In January 2025, Google began testing Ask for Me as a Search Labs experiment for nail salons and auto repair shops, according to Search Engine Journal. Google's agentic calling post describes US shoppers asking Google to call nearby stores to confirm a product is in stock. Each call reportedly opens by saying it is an automated system calling from Google on behalf of a user, according to Android Police. It does the same job Google Duplex was demonstrated doing at Google I/O in May 2018. In that demo, Google Assistant phoned a restaurant and a salon, TechCrunch reported.

Similarly, Meta launched public beta testing for business calling in September 2026 for its assistant Muse, as detailed in TechCrunch. We cover assistant architecture in our Meta Muse overview and detail commercial tiers in our review of Meta Muse subscription options. Concurrently, Instinct launched Instinct Concierge for restaurant and service reservations, which we explain in our Instinct agent overview. According to Reuters reporting on BNN Bloomberg, 404 Media, and TechSpot, Meta tested a "human concierge" in which trained contractors in call centers placed some Muse calls. Internal tests reportedly found those human-placed calls succeeded 95% to 98% of the time. Meta rolled the human-placed calls back. A Meta vice president said starting them without proper disclosure was "a miss."

These delegated assistant calls differ fundamentally from telemarketing campaigns because an individual customer initiates each inquiry. Federal robocall rules were written for robocall campaigns. We have not verified any court or FCC ruling on consumer-agent calls, so this use has not been tested. Google's Business Profile help page lists four ways to stop Google's calls. You can say "Please stop calling my business" on the call or turn off "Google automated calls and text messages" in your Business Profile's Advanced settings. You can reply STOP to its texts. To stop calls that map your phone tree, leave a voicemail with your business name and number at +1-650-206-5555. Meta has not published an official opt-out mechanism for Muse, so staff should implement our protocol for handling Meta Muse calls.

Proposed Federal Disclosure Mandates and State Regulations

In August 2024, the FCC issued a formal Notice of Proposed Rulemaking, recorded in the Federal Register publication. The proposal would define an "AI-generated call" and require callers to disclose that they use AI.

As of September 27, 2026, we have not verified that the FCC finalized this proposal. You should treat it as proposed. Several states have passed or proposed their own AI-call disclosure rules. Ask a lawyer licensed in your state what applies to you.

Call Recording Laws and Evidence Collection

Documenting illegal synthetic calls requires careful attention to jurisdictional recording statutes. Audio of an illegal AI sales call can support a complaint, but whether you may record the call depends on your state's consent rule.

State surveillance laws govern whether you can record telephone conversations. Certain jurisdictions allow recording when one participant approves. Other regions require all parties on the line to consent before capturing audio. Review our two-party consent state recording guide to confirm your local legal requirements before saving audio logs.

When collecting evidence without audio recordings, retain comprehensive digital documentation. Preserve call logs showing incoming caller ID data. Note the exact arrival timestamp, document the synthetic nature of the caller, and archive any automated voicemail messages delivered to your inbox.

Operational Responses and Practical Next Steps

Identifying automated synthetic voices requires recognizing specific conversational patterns. AI systems often display subtle pauses before responding, maintain an identical acoustic timbre across shifting conversational contexts, or struggle when confronted with sudden interruptions or ambiguous clarifying questions. You can review specialized acoustic analysis software in our audio detection tool comparison. Explore diagnostic techniques in our guide on identifying synthetic voices.

If an AI voice asks for payment, card details or passwords, hang up and call back on a number you already have for that company. Google says its automated calls never ask for payment, credit card information or account passwords. You can cross-reference fraudulent calling patterns through our breakdown of spam call classifications.

Organizations that plan outbound automated campaigns should not rely on receiving-side frameworks. Outbound marketing teams require dedicated legal counsel. They also need automated compliance controls. Our verdict on AI call legality would change if the FCC finalizes its proposed AI disclosure rules, or if a court or the FCC rules on calls placed by a customer's own AI agent. This is general information, not legal advice, so check with a lawyer about your own situation.

• Ask the caller directly whether you are speaking with an artificial intelligence or a live human agent.
• Hang up and dial back the company directly using a verified telephone number from an official invoice or website.
• Review your incoming telephone records today and document timestamps for any unsolicited synthetic voice contacts.