Back to Blog
AI RECEPTIONIST

Is it illegal to have AI call people?

Compliance And Data Security > GDPR Practices16 min read

Is it illegal to have AI call people?

Key Facts

The Short Answer: AI Calls Aren't Banned, But They're Heavily Regulated

So, is it illegal to have AI call people? No, not outright. But if you're using an AI-generated voice for marketing calls to mobile phones without prior express written consent, you're breaking federal law. That distinction matters, and it's where most businesses get into trouble.

The turning point came in February 2024, when the FCC unanimously adopted a Declaratory Ruling that AI-generated voices count as "artificial or prerecorded" voices under the Telephone Consumer Protection Act. In plain terms, your conversational AI is legally a robocall, no matter how natural it sounds.

Here's the part that surprises people: the sophistication of the technology makes no legal difference. As one TCPA law firm puts it, "The legal standard doesn't depend on how convincing the voice sounds. What matters is whether a live person is speaking to you in real time or whether technology is generating the words you hear." A system that answers questions, adapts its tone, and books appointments in real time still falls under the same rules as a canned robocall.

Why does this matter financially? Because the TCPA carries statutory damages of $500 to $1,500 per call, available without the caller proving any actual harm, and up to the full $1,500 if the violation was willful. You don't need a regulator to come after you, either. Private plaintiffs and class actions drive most TCPA enforcement, and the statute of limitations runs four years, meaning every noncompliant call in your logs is a live claim waiting to be filed. Run a modest campaign of a few thousand calls and the math stops being a technicality fast.

A few things worth keeping straight:

  • AI voices are treated as artificial under the TCPA regardless of how human they sound, per the FCC's February 2024 ruling.
  • Hybrid systems are covered too. If a dialer or recording initiates the call, even one where a live agent joins later, the prerecorded-message rules apply.
  • B2B isn't a safe harbor. Business calls face the same TCPA wireless restrictions as consumer calls.
  • Narrow exceptions exist for informational calls like appointment reminders, emergency messages, and HIPAA-compliant healthcare communications, as long as no marketing content is mixed in.

That last point is why compliant use cases are very real. A dental practice using an AI receptionist to confirm appointments, or a law firm following up on a lead who already opted in, sits on much firmer ground than a company blasting cold marketing calls. The law targets unconsented marketing, not the technology itself.

If you're weighing whether an AI phone system fits your business, the smart move is to book a free setup call and talk through consent, disclosure, and opt-out handling before you dial anyone. Consent, not the AI, is what determines legality. Get that right and the technology is yours to use.

If you're using AI to place outbound calls, three legal obligations do most of the heavy lifting: getting proper consent up front, honoring opt-outs quickly, and telling people they're talking to a machine. Miss any of them, and each AI call can cost you $500 to $1,500 in statutory damages under the TCPA, with a four-year window for claims and strict liability, meaning no proof of actual harm is required.

Obligation one: prior express written consent for marketing calls to mobile phones. The FCC's February 2024 Declaratory Ruling confirmed that AI-generated voices count as "artificial or prerecorded" calls, so they can't dodge TCPA coverage no matter how convincing they sound. Even conversational systems that respond to questions in real time require consent, and hybrid setups where a live agent joins later are still covered if technology initiated the call. Written consent must clearly authorize artificial-voice calls, specify the phone number, and can't come from pre-checked boxes.

Obligation two: honor revocation fast. The compliance center of gravity has shifted from acquiring consent to losing it. The Eleventh Circuit vacated the one-to-one consent rule in January 2025, but the revocation rules that took effect April 11, 2025 are very much alive. Under these rules, a consumer can revoke consent through any reasonable means, including saying or texting "stop," "cancel," "end," "revoke," or "unsubscribe," and you have 10 business days to honor it. As compliance analysts put it, if the message is clear, it counts, and the burden falls on you to prove otherwise. One confirmation text is allowed within five minutes, and it can't contain marketing content.

Obligation three: disclose AI status. AI agents that fail to identify themselves as non-human violate FTC deceptive practice guidelines. That's separate from the TCPA, and it's why best-practice call scripts identify the company, state the call's purpose, disclose that AI is on the line and that the call may be recorded, and offer an easy opt-out. The FCC has proposed going further with a formal in-call AI disclosure requirement, so this obligation will likely tighten.

Two more dates deserve a spot on your calendar:

  • The FCC's proposed AI-specific rules would define "AI generated call," require specific consent for AI calls, and mandate disclosure mid-call that AI is being used.
  • A universal cross-channel opt-out takes effect April 11, 2026, potentially letting a single "STOP" apply to all automated communications, including appointment reminders.

The practical takeaway for a small service business: informational and transactional calls like appointment confirmations don't require written consent as long as no marketing content sneaks in, which is exactly how a compliant AI receptionist setup should work. Answrr configures call scripts with disclosure language and opt-out handling built in from day one, so if you want the compliance groundwork handled for you, you can book a free setup call. The rules above still apply to any marketing calls you make separately, and state laws, from Florida's broad autodialer definition to New York's three-second opt-out requirement, often add stricter layers on top.

The good news is that you don't always need written consent to let AI place calls. Federal law carves out a handful of narrow exceptions, and they map neatly onto what service businesses actually do day to day: reminders, confirmations, and urgent messages.

The biggest exception covers purely informational and transactional calls. Appointment reminders, confirmations, and package delivery notifications don't require prior express written consent, as long as the call contains no marketing content, according to legal analysis of TCPA consent rules. That's why an AI receptionist handling reminders and confirmations fits naturally inside the law. The moment you slip a promotional line into the message, you've left the exception behind.

Healthcare communications get their own carve-out, but it comes with strict conditions. The HIPAA exemption covers treatment-related messages only, and each call must meet every requirement: no promotional content, a number the patient provided themselves, a voice message under one minute, calls placed between 8 a.m. and 9 p.m. in the recipient's time zone, and opt-outs honored immediately, per a TCPA compliance guide. Miss one condition and the exemption fails entirely.

Other permitted categories include emergency and public safety messages, fraud alerts from financial institutions, and prerecorded telemarketing to residential landlines (which still must respect Do Not Call rules and offer an automated opt-out).

  • Appointment reminders and confirmations with zero marketing content
  • Emergency and public safety messages
  • Fraud alerts from financial institutions
  • HIPAA-compliant, treatment-related healthcare communications
  • Prerecorded telemarketing to residential landlines, with DNC compliance

Two warnings before you lean on these exceptions. First, B2B is not a safe harbor. Business calls face the same TCPA wireless restrictions as consumer calls, and "business call" is not a universal exemption, as industry guidance on AI outbound calls points out. Calling a plumber's cell phone with an AI voice is not automatically legal just because it's a business relationship.

Second, state law can be stricter than federal law. Florida's "Mini-TCPA" uses an extremely broad definition of an autodialer, and New York requires an opt-out opportunity within three seconds of a call's start, regardless of consent. Roughly half of states have their own calling-time rules, and recording consent requirements vary by state. If you operate across state lines, audit each jurisdiction before you dial.

If you'd rather have someone configure compliant reminder and confirmation calling for you, Answrr sets up its AI receptionist white-glove, and you can book a free setup call to walk through your use case.

Your Compliance Playbook for AI Calling

The FCC's February 2024 ruling made one thing clear: AI voices count as "artificial" calls under the TCPA, and that changes how you have to run outbound calling. The good news is that compliant AI calling is entirely achievable if you build the right safeguards into your process from day one.

Start with consent. Before any AI-generated marketing call to a mobile number, you need prior express written consent that clearly authorizes artificial-voice calls, specifies the phone number, and isn't buried in a pre-checked box or conditioned on a purchase. Since the FCC treats AI voices the same as prerecorded robocalls, per the FCC's Declaratory Ruling, how human the voice sounds makes no legal difference.

Next, build scripts that disclose what callers are hearing. FTC deceptive-practice guidelines require AI agents to identify themselves as non-human, and the FCC has proposed making in-call AI disclosure a formal requirement. Every script should do four things: identify your company, state the call's purpose, disclose that the caller is an AI, and offer an easy opt-out.

Opt-out handling is where enforcement has shifted. Under rules effective April 11, 2025, consumers can revoke consent through any reasonable means, and you have 10 business days to honor the request. That includes phrases like "stop," "cancel," "unsubscribe," or "quit," so many businesses now use AI tools that detect varied revocation language and process it automatically.

Then audit your state-level exposure. Florida's Mini-TCPA has an extremely broad autodialer definition, New York requires an opt-out opportunity within three seconds of the call's start, and roughly half the states have their own calling-time and call-recording consent rules. Federal compliance alone isn't enough.

Finally, keep marketing content out of reminder calls. Appointment reminders and confirmations fall under the informational/transactional exception and don't require written consent, but the moment promotional language creeps in, that protection evaporates. The healthcare exemption is even stricter: no promotional content, calls under one minute, and immediate opt-out honoring.

This is where an inbound-focused setup earns its keep. An AI receptionist answering incoming calls and placing appointment confirmations sits comfortably within the informational exception, which is why platforms like Answrr are built around answering and reminders rather than cold outbound campaigns. If you're weighing an inbound setup for your service business, you can book a free setup call to see how a compliant configuration works in practice.

One last caution: B2B is not a safe harbor. Business calls to wireless numbers face the same TCPA restrictions as consumer calls, so apply the same consent discipline regardless of who you're dialing.

Voice Data Risks Beyond the TCPA: GDPR, HIPAA, and BIPA

The TCPA gets most of the attention, but it only covers one slice of the risk. If your AI calls capture voice recordings, generate voiceprints, or touch health information, you're also in GDPR, BIPA, and HIPAA territory, and those regimes carry penalties of their own.

GDPR regulators have issued guidance treating voice biometrics as special category data, which means the bar for lawful processing is higher than for ordinary customer data, according to voice AI compliance analysis. Mishandling voice data can trigger fines up to €20 million or 4% of global annual revenue, whichever is higher. If your AI system records callers in Europe or builds voice models from those recordings, this applies to you.

Illinois adds another layer. The state's Biometric Information Privacy Act (BIPA) covers voiceprints as biometric identifiers, so an AI phone system that captures and stores voice characteristics from Illinois residents can create liability even when the call itself was TCPA-compliant.

One trap catches businesses off guard more than any other: consent creep. Consent a caller gives for one purpose, like receiving an appointment reminder, does not cover training your AI model on that call recording. Regulators require separate, purpose-specific consent for AI training, per the same compliance guidance. A caller who agreed to a reminder call never agreed to become training data.

Medical practices face a third exposure. HIPAA penalties start at $100 per violation and can reach $1.5 million annually per violation category. The TCPA does carve out an exception for HIPAA-compliant healthcare calls, but the conditions are strict:

  • No promotional content of any kind mixed into the call
  • Calls only to numbers provided by the patient
  • Voice messages under one minute, texts under 160 characters
  • Calls placed between 8 a.m. and 9 p.m. in the recipient's time zone
  • Opt-outs honored immediately

Fall outside those conditions, and you've lost both the TCPA exemption and your HIPAA safe harbor in one call.

This is why the vendor you pick matters as much as the technology itself. A platform like Answrr that handles compliance setup for you, from consent language to call handling rules, removes the guesswork that turns a helpful tool into a legal liability. If you want an AI receptionist that answers every call 24/7 without creating regulatory exposure, book a free setup call and let the compliance work happen before your first call goes out, not after your first complaint.

Frequently Asked Questions

Is it actually illegal to have AI call people, or just regulated?
AI calling isn't banned outright, but the FCC's February 2024 Declaratory Ruling made clear that AI-generated voices count as "artificial or prerecorded" calls under the TCPA. That means AI marketing calls to mobile phones require prior express written consent, and violations carry statutory damages of $500 to $1,500 per call. The law targets unconsented marketing, not the technology itself.
Does it matter if my AI sounds really human and can hold a real conversation?
No, the sophistication of the technology makes no legal difference. As TCPA legal analysis puts it, what matters is whether a live person is speaking or technology is generating the words, so even conversational AI that answers questions and books appointments is legally a robocall. Hybrid systems where a live agent joins later are also covered if technology initiated the call.
Do I need written consent for AI appointment reminders and confirmations?
No. Purely informational and transactional calls like appointment reminders, confirmations, and package delivery notifications don't require prior express written consent, as long as no marketing content is mixed in, per legal analysis of TCPA consent rules. This is exactly why an AI receptionist handling reminders and confirmations fits naturally within the law. The moment a promotional line slips in, the exception evaporates.
What happens if someone says 'stop' during an AI call, how fast do I have to honor it?
Under rules effective April 11, 2025, consumers can revoke consent through any reasonable means, saying or texting "stop," "cancel," "end," or "unsubscribe", and you have 10 business days to honor it, with compliance analysts noting the burden is on you to prove the message wasn't a valid revocation. One confirmation text is allowed within five minutes, and it can't contain marketing content.
Are B2B calls exempt from these AI calling rules?
No, B2B is not a safe harbor. Business calls face the same TCPA wireless restrictions as consumer calls, and as industry guidance on AI outbound calls points out, "business call" is not a universal exemption. Calling a plumber's cell phone with an AI voice isn't automatically legal just because it's a business relationship.
Do I have to tell people they're talking to an AI?
Yes. AI agents that fail to identify themselves as non-human violate FTC deceptive practice guidelines, and the FCC has proposed a formal in-call AI disclosure requirement. Best-practice scripts identify your company, state the call's purpose, disclose that AI is on the line, and offer an easy opt-out, which is why platforms like Answrr build disclosure language and opt-out handling into their AI receptionist setups from day one. If you want that groundwork handled for you, book a free setup call to review your use case before dialing anyone.

The Bottom Line: Consent, Not Technology, Decides What's Legal

AI calling isn't banned, it's regulated, and the distinction matters more than most businesses realize. The FCC's February 2024 ruling put AI-generated voices squarely under the TCPA, meaning marketing calls to mobile phones require prior express written consent, opt-outs must be honored within 10 business days, and callers must disclose they're speaking with a machine. Violations run $500 to $1,500 per call with a four-year statute of limitations, and B2B offers no safe harbor. The good news: informational and transactional calls, appointment reminders, confirmations, and HIPAA-compliant healthcare messages, sit on much firmer legal ground, provided no marketing content creeps in. That's precisely where an AI receptionist earns its keep, answering every call 24/7 and handling reminders without venturing into cold outbound territory. If you'd rather build compliance in from day one than audit it after a complaint, book a free setup call and let Answrr configure your scripts, disclosures, and opt-out handling before your first call goes out.

Get AI Receptionist Insights

Subscribe to our newsletter for the latest AI phone technology trends and Answrr updates.

One call. Set up by us. Live in 24 hours.

Ready to stop missingrevenue calls?

Start Your Free Trial
Free setup call included
Plans from $49/mo
No credit card to book
Is It Illegal to Have AI Call People? TCPA Rules Explained