Are AI Voice Scam Calls Illegal? FCC and State Law (2027)

Yes. As of November 2026, an AI voice scam call breaks federal law in more than one way: FCC Declaratory Ruling 24-17 applies the Telephone Consumer Protection Act (TCPA), 47 U.S.C. 227, to AI-generated voices, so the call needs the called party's prior express consent absent an emergency purpose or exemption, and using the call to defraud someone of money can be charged as wire fraud under 18 U.S.C. 1343. Some states add crimes of their own, including Maryland, Arizona and Pennsylvania.

Status as of November 5, 2026: The FCC ruling has been in effect since February 8, 2024. The FTC's Impersonation Rule covers government and business impersonators, and the FTC says it "is still evaluating its options" on individuals. We found no final FCC rule requiring callers to disclose AI, and no federal statute that makes AI fraud a separate crime. The federal AI scam bills we checked have not moved past referral to committee.

This page covers scam calls and robocalls under federal law, with a sample of state statutes. Video and business deepfake fraud is on our deepfake fraud and impersonation page. Rights in a person's voice are on our AI voice cloning laws page.

Are AI Voice Scam Calls Illegal?

Yes, and several federal laws reach them, each doing a different job. One regulates the call, one regulates the caller ID, and others reach the theft. AI is not an element of the fraud statutes, so the same scheme can be charged whether a machine or a person did the talking.

The TCPA also reaches overseas callers on its face. Section 227(b)(1) applies to "any person within the United States, or any person outside the United States if the recipient is within the United States." Finding that person is a separate problem.

FCC 24-17: AI Voices Under the TCPA

The FCC adopted Declaratory Ruling 24-17 in CG Docket 23-362 on February 2, 2024, and released it on February 8, 2024, effective on release. This is the ruling people mean by "the FCC AI robocall rule." Its central line, from paragraph 2:

"In this Declaratory Ruling, we confirm that the TCPA's restrictions on the use of 'artificial or prerecorded voice' encompass current AI technologies that generate human voices."

The ruling goes on to say that such calls "require the prior express consent of the called party to initiate such calls absent an emergency purpose or exemption." If a robocall introduces an advertisement or contains telemarketing, the FCC's rules require prior express written consent. Our TCPA overview explains how consent works.

The ruling covers technologies that "either wholly simulate an artificial voice or resemble the voice of a real person taken from an audio clip to make it appear as though that person is speaking on the call." So a generic synthetic voice and a clone of a real person are both covered. It also points to an existing rule, 47 CFR 64.1200(b)(1), under which every artificial or prerecorded voice message must state clearly, at the beginning, who is responsible for the call.

Two limits matter. The ruling reads the TCPA; it is not a ban on AI voices or on voice cloning. And the statute has exceptions: 47 U.S.C. 227(b)(1)(B) excepts residential calls for emergency purposes, calls made solely to collect a debt owed to or guaranteed by the United States, and calls the FCC exempts.

How much weight courts give FCC 24-17

An FCC ruling is the agency's reading of the statute, and it does not bind a federal trial court in a private suit. In McLaughlin Chiropractic Associates, Inc. v. McKesson Corp., 606 U.S. 146 (2025), a TCPA case, the Supreme Court said: "The District Court is not bound by the FCC's interpretation of the TCPA."

The Court told the district court to interpret the statute "under ordinary principles of statutory interpretation, affording appropriate respect to the agency's interpretation." So a court hearing a private suit over an AI voice call decides for itself what "artificial" means.

The FCC's AI disclosure proposal, FCC 24-84

In August 2024 the FCC proposed going further. FCC 24-84, adopted August 7 and released August 8, 2024, proposed to define "AI-generated call" and to require callers using AI-generated voice to say so at the beginning of each call.

As of November 5, 2026, it remains a proposal: one search limited to FCC and Federal Register sites found no final rule. The disclosure duty it proposes should not be read as law.

Can You Sue Over an AI Voice Scam Call?

The TCPA gives a person who receives an unlawful call a private claim. Under 47 U.S.C. 227(b)(3), a person may, "if otherwise permitted by the laws or rules of court of a State," sue to stop the calls, "to recover for actual monetary loss from such a violation, or to receive $500 in damages for each such violation, whichever is greater," or both.

If the court finds the defendant "willfully or knowingly" violated the law, it "may, in its discretion, increase the amount of the award" to no more than three times that amount. That is $1,500 per violation when the base is $500. The general federal deadline is 4 years after the claim accrues, under 28 U.S.C. 1658(a), unless another law sets a different one.

The practical limit is finding someone who can pay. Spoofed numbers and overseas callers make that hard, and a judgment is worth only what can be collected. The wire fraud and extortion statutes give a victim no private right to sue.

State officials have their own route. Under 47 U.S.C. 227(g)(1), a state attorney general who has reason to believe someone is engaged in a pattern or practice of unlawful calls to the state's residents may bring a civil action on their behalf.

How damages are counted and how class actions work are on our TCPA damages and lawsuits page.

The AI Biden Robocalls: Lingo Telecom and Steve Kramer

The best-known AI voice robocall case is not a money scam. On January 21, 2024, two days before the New Hampshire primary, potential voters received calls carrying an AI-generated message made to sound like President Biden. The FCC's order calls it the "Deepfake Message." Neither FCC action below rests on the TCPA's artificial voice provision.

Lingo Telecom: $1,000,000 consent decree. Lingo, a voice service provider, completed 3,978 of the calls and signed them with A-level attestations under the STIR/SHAKEN caller ID framework. The FCC Enforcement Bureau resolved the matter under 47 CFR 64.6301(a) in Consent Decree DA 24-790, adopted August 21, 2024.

Lingo agreed to pay "a civil penalty to the United States Treasury in the amount of one million dollars ($1,000,000)" and to adopt a compliance plan. Under the decree it "may only apply an A-level attestation to a call if Lingo Telecom itself has provided the Caller Identity" to the calling party.

Steve Kramer: $6,000,000 FCC forfeiture. In Forfeiture Order FCC 24-104, adopted September 26 and released September 30, 2024, the FCC held that "Kramer is liable for a forfeiture of $6,000,000." The order says 9,581 calls went out with a spoofed caller ID number, and the forfeiture rests on a verified subset of 3,000 of them.

The basis is the Truth in Caller ID Act, 47 U.S.C. 227(e). The FCC said such a violation "is not dependent on the content of the call," though AI voice-cloned messages paired with misleading caller ID "may further demonstrate a caller's intent to defraud, cause harm, or wrongfully obtain something of value." So this was a forfeiture for spoofing, not a fine for deepfaking. The order notes Kramer did not respond to the earlier notice, and it directs payment within 30 calendar days.

Kramer was not convicted. A New Hampshire jury acquitted him of all state charges in June 2025, the Associated Press reported: 11 felony voter suppression charges and 11 candidate impersonation charges. He told the AP he won't pay the FCC fine. We did not find whether it has been collected.

FTC Impersonation Rule: Banks and Agencies Yes, Grandchildren No

The FTC's Rule on Impersonation of Government and Businesses, 16 CFR Part 461, took effect on April 1, 2024. Section 461.2 bars materially and falsely posing, "directly or by implication," as "a government entity or officer thereof." Section 461.3 does the same for "a business or officer thereof." An AI voice that claims to be your bank or a federal agency can fall under it, and the FTC says the rule allows civil penalties against violators.

The rule does not cover a scammer who impersonates a private person, such as a grandchild. The FTC proposed extending it to individuals, and that has not been adopted. In a notice published October 1, 2026 (Federal Register document 2026-20143), the FTC said: "With respect to the impersonation of individuals provision, the Commission held an informal hearing in January 2025 and is still evaluating its options."

The FTC also stepped back from reaching companies that supply the tools. In December 2024 it said it "has decided not to proceed with the SNPRM's proposed means and instrumentalities provision at this time." The October 2026 notice opens a narrower question, limited to "ad-optimization practices by digital marketplace platforms." Comments close November 30, 2026.

Federal Fraud Statutes That Reach a Cloned Voice Call

These statutes can be charged against an AI voice scam. None mentions AI, so this section describes what the text reaches, not a record of prosecutors using it against cloned voices. Prosecutors bring these charges; a victim cannot.

The voice itself is an open question. Section 1028(d)(7)(B) lists "unique biometric data, such as fingerprint, voice print, retina or iris image, or other unique physical representation." Whether a synthetic imitation of a voice counts as a "voice print" is not settled by the text. A scam that also uses a real person's name is on firmer ground, because the definition covers a name.

Dubin limits Section 1028A. In Dubin v. United States, 599 U.S. 110 (2023), the Supreme Court held that a defendant uses another person's means of identification "in relation to" a predicate offense "when this use is at the crux of what makes the conduct criminal." In an impersonation scam the impersonation is usually the deception itself, but we did not find a case applying that test to a cloned voice.

We found no federal statute that makes AI fraud a separate crime. Of the bills listed in the 2027 section below, the AI Fraud Accountability Act would create a new federal crime; the others rely on FTC enforcement or on disclosure and penalty changes to the TCPA.

Grandparent Scam AI Calls: Which Law Fits

The grandparent scam is the same old fraud with a better voice. The FCC named the risk in its ruling: voice cloning "can convince a called party that a trusted person, or someone they care about such as a family member, wants or needs them to take some action that they would not otherwise take."

The FBI warns that criminals "generate short audio clips containing a loved one's voice to impersonate a close relative in a crisis situation, asking for immediate financial assistance or demanding a ransom."

The federal fit is wire fraud, plus the TCPA if the voice was artificial and there was no consent. The FTC rule does not fit unless the caller also poses as a government entity, an officer or a business, such as a fake police officer.

State law depends on whose harm the statute names. Arizona's and Maryland's texts appear to fit a call that fools a grandparent. New Hampshire's fits less cleanly, because its purpose element points to harm "to the identifiable person," the one who was faked, not the one who was fooled. That is our reading of the text; we found no court decision on it.

State Laws on AI Voice Fraud: Maryland SB 8 and Others

Every state has general fraud and theft laws that can apply to a scam call, however the voice was produced. The table below is a sample of statutes that name AI, deepfakes or computer-generated voices. It is not a 50-state survey, and a state left out may still have a relevant law. State-by-state coverage is on the deepfake laws hub.

Maryland SB 8. The act was approved May 12, 2026, and took effect October 1, 2026. It bars knowingly, willfully and with fraudulent intent using artificial intelligence or a deepfake representation to "cause harm to another by impersonating, falsely depicting, or claiming to represent that person or any other person." It separately bars using them to create or distribute false records with listed intents.

"Deepfake representation" includes an audio recording "that is indistinguishable from an actual and identifiable human being." "Harm" means physical injury, serious emotional distress or economic damages, and a sentence may run consecutive to the sentence for the underlying crime. A victim may also bring a civil action, and a court may issue an injunction. A web search found no reported appellate decision applying the section yet.

Arizona's carve-outs. The statute defines "defraud" as "to make a false representation or material omission to deceive another person to gain a benefit." Comedy, parody, artistic expression, criticism and circumstances "where it is clear to a reasonable listener or viewer" that the recording was digitally manipulated are not fraudulent.

Keep the verbs apart. New Jersey's offense can be complete at generating the media, if that is for a criminal purpose, and the statute separately reaches soliciting, disclosing or using it for that purpose. Pennsylvania needs both creating and distributing. Washington's offense is distributing. Arizona's is using.

Civil claims differ too. New Hampshire's RSA 507:8-j lets the person faked sue for damages, and New Jersey lets a victim sue, with liquidated damages of at least $1,000 for each knowing or reckless violation.

Laws that are not stand-alone fraud crimes. California's SB 1111 (Ch. 862, approved September 30, 2026) adds Penal Code 540, so that where false impersonation is an element of a crime, it "includes the use of a digital replica with the intent to impersonate another." That defines an element of existing crimes; it is not a new crime. Utah Code 76-2-107 says an actor may be found guilty of an offense committed with the aid of generative AI, which is a liability rule, not a fraud offense.

Virginia's HB 2124 was signed in 2025, but conditioned on reenactment by the 2026 session, and it does not appear in the Code. More on video and business fraud is on our deepfake fraud and impersonation page.

Is AI Voice Cloning Illegal by Itself?

Not under the federal laws on this page. The FBI puts it this way: "The creation or distribution of synthetic content is not inherently illegal; however, synthetic content can be used to facilitate crimes, such as fraud and extortion." That sentence is not a statement that every use is lawful.

We found no federal law that bars cloning a voice. The TAKE IT DOWN Act does not fill the gap, because it reaches intimate visual depictions, not audio. State law can differ: some states have civil laws reaching digital replicas or simulated voices, and New Jersey's text reaches generating deceptive media for a criminal purpose.

Our AI voice cloning laws page and right of publicity page cover what a person whose voice was copied can do.

What to Do If You Get an AI Voice Scam Call

Do not rely on the voice. The FBI says it can be difficult to identify when content is AI-generated, and its advice for these calls is practical.

• Hang up and verify. Call the person or organization back on a number you looked up yourself. The FBI says to hang up, research the contact and "call the phone number directly."
• Agree on a family secret word. The FBI recommends "a secret word or phrase with your family to verify their identity."
• Do not pay a stranger. The FBI says not to send "money, gift cards, cryptocurrency, or other assets" to people you do not know or have met only online or over the phone.
• Keep the details. Note the date, time, phone number, what was said and how you were told to pay. The FBI asks for this kind of detail in a report. If you want to record a suspicious call, rules differ by state; see recording laws by state and the federal Wiretap Act.

Then report it. Each office handles a different piece.

• FTC: ReportFraud takes reports of the scam itself.
• FBI Internet Crime Complaint Center: file at ic3.gov. The FBI asks for identifying details, transaction records and how contact began.
• FCC: file a complaint at consumercomplaints.fcc.gov about the robocall, spoofed caller ID or unwanted AI voice call.
• Your state attorney general: find yours at usa.gov. A consumer protection office may also help, listed at usa.gov/state-consumer.
• Your bank: call the number on your card or statement right away if any money moved.
• Police: call local police if the caller threatened anyone or you are worried a relative really is in danger.

For advice on your own loss, your state bar's lawyer referral service or a local legal aid office can point you to a lawyer. If the target is an older relative, our elder fraud guide covers warning signs and how to report.

If You Sent Money: Bank and Regulation E Limits

Regulation E, 12 CFR Part 1005, turns on who started the transfer. An unauthorized transfer is one "initiated by a person other than the consumer without actual authority to initiate the transfer and from which the consumer receives no benefit."

If a scammer uses a stolen login or code to move money, federal rules generally treat that as unauthorized, and your liability is limited if you report quickly. These protections cover electronic transfers from an account held primarily for personal, family or household purposes. The CFPB says that where a caller tricks a consumer into giving login information or a texted code, "the transfer is an unauthorized EFT under Regulation E."

Under 12 CFR 1005.6(b), liability is generally capped at the lesser of $50 or the amount taken before notice if you tell the bank within two business days after learning your login or code was taken. After that, the cap generally rises to the lesser of $500 or a computed sum. Also report any unauthorized transfer within 60 days after the bank sends the statement showing it, because you can be liable for later transfers that notice in that window would have stopped, and that amount is not held to the $500 cap.

The rule does not clearly cover a payment you sent yourself, even if you were deceived. We found nothing in the rule or the CFPB guidance we read that addresses a transfer you start yourself after a fake voice fools you, so do not assume a bank must refund it.

The CFPB's suit against Early Warning Services and three banks over Zelle, filed December 20, 2024, was dismissed with prejudice on March 5, 2025, according to the CFPB.

Regulation E excludes transfers through Fedwire or a similar wire system used primarily between financial institutions or businesses (12 CFR 1005.3(c)(3)), so these limits may not apply to a wire. This page does not cover wire transfers or card chargebacks in detail. Our guide on how to get money back after a scam goes through each payment method.

AI Scam Complaints in the FBI's IC3 Data

The FBI's IC3 report for 2025 lists 1,008,597 complaints and $20.877 billion in losses, a 26% increase in losses from 2024. Complaints carrying its "AI Related" descriptor numbered 22,364, with $893,346,472 in adjusted losses.

Read those numbers with care. The descriptor means the report "contains a reference to artificial intelligence (AI)," so these are self-reported complaints that referenced AI, not proven AI crimes. On grandparent or distress scams, where the IC3 says voice cloning is used "to mimic the sound of a loved one in distress," it reports: "Victims claimed losses over $5 million in 2025."

What Changes for AI Voice Scam Calls in 2027

These are the dated events we found as of November 5, 2026.

• November 30, 2026: comments close on the FTC's notice, Federal Register document 2026-20143. It covers platform ad practices. The individuals provision stays under evaluation, with no date given.
• January 1, 2027: California's SB 1111 takes effect by the default date for non-urgency laws, adding Penal Code 540. That date is our reading of Cal. Const. art. IV, sec. 8(c), not a date we read in the act.
• January 1, 2027: Nebraska's LB838 becomes operative. Neb. Rev. Stat. 87-302(a)(26) requires a social media platform that accepts payment for ads to have identity verification for advertisers and an unlawful impersonation detection and mitigation program. It is a deceptive trade practice provision, and Neb. Rev. Stat. 87-303.08 makes a violation of the Uniform Deceptive Trade Practices Act a Class II misdemeanor except as otherwise provided in the act.
• January 3, 2027: the 119th Congress ends. A bill that is not enacted before the 119th Congress ends on January 3, 2027 does not carry over. It would have to be introduced again in the next Congress. Congress could still act before then.

Bills in that position, all at referral stage as of November 5, 2026, include the Artificial Intelligence Scam Prevention Act (S. 3495), the AI Fraud Accountability Act (S. 3982, H.R. 7786) and the QUIET Act (H.R. 1027, S. 3354), which would make AI robocalls disclose AI at the start. None is law. Our federal deepfake bills page tracks them.

Two items have no date: the FCC's AI disclosure proposal and the FTC's individuals provision. We did not check bills prefiled for 2027 state sessions.

This page is general legal information about US law as of November 5, 2026, not legal advice. A lawyer licensed in your state can advise you on your situation.

Frequently Asked Questions

Are AI voice scam calls illegal?

Yes. As of November 2026, FCC Declaratory Ruling 24-17 treats AI voices as artificial or prerecorded voices under the TCPA, so the call needs prior express consent absent an emergency purpose or exemption, and using the call to defraud someone of money can be charged as wire fraud under 18 U.S.C. 1343. Some states, including Maryland and Arizona, add their own crimes.

Did the FCC ban AI robocalls?

Not outright. FCC 24-17 says AI voices count as artificial or prerecorded voices under the TCPA, so the calls need the called party's prior express consent absent an emergency purpose or exemption. Robocalls with telemarketing need prior express written consent. The FCC's August 2024 proposal for an AI disclosure rule is still a proposal; we found no final rule.

Is AI voice cloning illegal?

Not by itself under the federal laws on this page, and we found no federal law that bars cloning a voice. The FBI says creating or distributing synthetic content 'is not inherently illegal,' though it can be used to commit crimes. Using a clone to defraud someone can be charged as a crime, and some state laws reach digital replicas or forged likenesses.

Does the FTC Impersonation Rule cover a scammer pretending to be my grandchild?

No. As of November 5, 2026, the rule in 16 CFR Part 461 covers impersonating government entities and businesses. The FTC says it is still evaluating its options on the individuals provision. The scam can still be charged as wire fraud.

Was Steve Kramer convicted over the AI Biden robocalls?

No. A New Hampshire jury acquitted him of all state charges in June 2025, the Associated Press reported. Separately, the FCC held him liable for a $6,000,000 forfeiture under the Truth in Caller ID Act in FCC 24-104. He told the AP he won't pay it, and we did not find whether it has been collected.

Was anyone fined for the AI Biden robocalls?

Yes. The FCC imposed the $6,000,000 forfeiture on Steve Kramer for spoofed caller ID, not for the AI voice itself. Lingo Telecom, which completed 3,978 of the calls, agreed to a $1,000,000 civil penalty and a compliance plan in Consent Decree DA 24-790.

Can you go to jail for an AI voice scam?

Yes. Federal wire fraud carries up to 20 years, or up to 30 years if a financial institution is affected. Aggravated identity theft can add 2 years when another person's means of identification is misused, though how it applies to a cloned voice is not settled. State crimes may also apply, such as Maryland's felony with up to 5 years, or 10 with two or more victims.

Can I sue someone who called me with an AI voice?

Possibly. The TCPA lets a person who received an unlawful call sue for actual loss or $500 per violation, whichever is greater, and a court may raise that up to three times for willful or knowing violations. The general federal deadline is 4 years. The hard part is finding and collecting from the caller, since numbers are often spoofed and callers may be overseas.

Will my bank refund money I sent after an AI voice scam?

It depends on who moved the money. If a scammer used a stolen login or code, federal rules generally treat the transfer as unauthorized, and your liability is limited if you report quickly. That covers electronic transfers from a personal account, and a bank wire may fall outside it. Report an unauthorized transfer within 60 days after the bank sends the statement showing it, or you can be liable for later transfers. The rule does not clearly cover a payment you sent yourself, even if you were deceived. Call your bank right away either way.

How can I tell if a call uses a cloned voice?

Often you cannot be sure by listening. The FBI suggests listening closely to tone and word choice, but it also says to hang up, call back on a number you looked up yourself, and set a secret word or phrase with your family.

Can I record a call I think is an AI scam?

Under federal law, 18 U.S.C. 2511(2)(d) lets a party to a call record it unless the recording is made to commit a criminal or tortious act. Some states require every party's consent, so the rule depends on where you and the caller are.

Where do I report an AI voice scam call?

Report the scam to the FTC at reportfraud.ftc.gov and the FBI at ic3.gov, and contact your state attorney general. Report a robocall or spoofed caller ID to the FCC at consumercomplaints.fcc.gov. If money moved, call your bank first. Call local police if the caller threatened anyone or you think a relative may really be in danger.