Yes, AI Can Call — But the FCC Made One Thing Clear

Yes, you can use AI to call someone — but only with the right consent, and the FCC has made the rules unambiguous. In February 2024, the Commission issued a Declaratory Ruling confirming that AI-generated voices are "artificial or prerecorded voice" under the TCPA, meaning every AI voice call requires prior express consent of the called party. The ruling explicitly closed the "live agent equivalent" loophole: the TCPA "does not allow for any carve out of technologies that purport to provide the equivalent of a live agent."

• Marketing AI calls require prior express written consent in most states
• Informational calls (reminders, notifications) can proceed on prior express consent, which may be oral or implied
• An existing business relationship does not exempt AI calls — a live agent may call a past customer on the DNC list; an AI agent may not without separate consent
• The caller bears liability, not the platform or lead seller

The stakes are severe. TCPA statutory damages run $500–$1,500 per call with no aggregate cap, and class-action filings have surged 95% year over year with aggregate verdicts exceeding $925 million. Recent settlements have reached $5M–$20M ranges, including a $9.95 million settlement by Gen Digital and a $19 million settlement by QuoteWizard tied to vendor-chain consent failures. Platforms like HighLevel explicitly do not perform contact-level consent validation — the business placing the call is responsible.

My AI Call Center treats this as a pre-launch discipline: list source and consent records are reviewed before any campaign runs, and lists without clear permission records are flagged or declined. Consent must be documented and retained for at least four years (seven recommended), including the exact language shown, timestamps, and proof of human interaction. Disclosure, opt-out handling, and revocation honoring within 10 business days are built into every call, alongside state-specific calling windows and recording rules. Campaign requirements vary by location, industry, contact type, and consent status — clients are responsible for obtaining appropriate legal guidance before launch.

The Consent Tier Your Call Actually Requires

Not every AI call needs the same paperwork. The consent you need depends entirely on what the call is trying to accomplish — and getting the tier wrong is where most AI calling programs run into trouble.

Since the FCC's February 2024 Declaratory Ruling classified AI-generated voices as "artificial or prerecorded voice" under the TCPA, every AI voice call requires prior express consent at minimum. But "prior express consent" is only the floor. The purpose of the call determines how much higher you need to go.

Calls that remind, confirm, notify, or update — appointment reminders, delivery notifications, fraud alerts, payment reminders — fall into the informational or transactional tier. According to TCPA compliance guidance for voice AI, these calls can proceed on prior express consent, which may be oral or even implied.

Implied consent has limits, though. As DNC.com's consent breakdown explains, when a consumer knowingly provides their number in the normal course of business, consent is implied only for messages that "closely relate" to the purpose for which the number was given. A patient who gives a clinic their number for appointment reminders has not consented to upsell calls.

The moment a call's purpose includes encouraging a purchase — a renewal quote, an upsell, a win-back offer — it becomes telemarketing, and the standard jumps to prior express written consent. Per the FCC definition summarized by DNC.com, valid written consent must contain four elements:

• A written agreement bearing the called party's signature (electronic signatures under E-SIGN qualify)
• The specific phone number to be called
• Clear authorization for telemarketing via autodialer or artificial/prerecorded voice
• A disclosure that consent is not a condition of purchase

ActiveProspect's consent guidance adds practical details: the language must identify the business or brand, sit near the submission button, and avoid pre-checked boxes. Vague "and our partners" language is a known weak point — courts are skeptical of it, and the defensible practice is re-consenting co-registered leads before any AI dialing.

Here is the trap: you cannot dress a marketing call in informational clothing. An "account check-in" that pivots to an upsell is a marketing call, full stop. Compliance analysts note that regulators assess the call's purpose, not its opening sentence — so a friendly first thirty seconds does not lower your consent tier.

The financial exposure makes this distinction worth taking seriously. TCPA statutory damages run $500 to $1,500 per call with no aggregate cap, and TCPA compliance guides point to class-action filings rising sharply year over year. A campaign that miscategorizes five thousand marketing calls as "informational" is not a gray area — it is a liability stack.

This is why consent verification has to happen before a single dial. At My AI Call Center, list source and consent records are reviewed before any campaign launches, and lists without clear permission records are flagged — or declined outright. Matching each campaign type to its correct consent tier is not a formality; it is the decision that determines whether the campaign is defensible at all.

Three Expensive Myths That Sink AI Calling Campaigns

Most AI calling campaigns don't fail because the technology underperformed. They fail because someone believed one of three myths about consent — and with TCPA statutory damages running $500 to $1,500 per call with no aggregate cap, a single mistaken assumption can turn a 10,000-call list into a seven-figure problem.

An existing business relationship exempts only live-agent calls from Do-Not-Call Registry restrictions. The moment an AI voice dials, the artificial voice itself triggers the consent requirement — no carve-out exists. The FCC was explicit that the TCPA "does not allow for any carve out of technologies that purport to provide the equivalent of a live agent," per the legal analysis of the February 2024 ruling.

One compliance playbook calls assuming EBR covers AI calls "the single most expensive misunderstanding in the AI outbound playbook" — a live agent may call a past customer on the DNC list, but an AI agent may not call that same person without separate consent, according to Retell AI's TCPA analysis.

Accepting an outbound calling platform's terms of service is not contact consent. HighLevel's own documentation states it plainly: the platform "does not perform platform-level contact consent validation," and businesses are responsible for ensuring contacts can legally receive AI calls, per HighLevel's compliance guidance. The platform even warns users not to treat a workflow terms checkbox as contact consent.

Liability follows the caller, not the vendor. Under the TCPA, the business making the calls — not the lead seller or the software — bears the liability if consent proves invalid, as ActiveProspect's consent guidance confirms. The QuoteWizard case, which ended in a $19 million settlement, is the reference example for failing to trace consent through the vendor chain, per the same compliance research.

Co-registration and bought "warm cold" lists have no legal meaning. Courts are skeptical of consent language buried in "and our partners" clauses, and the defensible practice is re-consenting every co-reg lead before any AI dialing, according to the research. Bot-submitted leads carry no valid consent at all, and consent attaches to the person, not the number — a reassigned number voids prior permission, per DNC.com's consent standards.

What defensible consent actually looks like:

• A signed agreement naming the business and the specific phone number
• Language authorizing telemarketing via artificial or prerecorded voice
• A clear statement that consent is not a condition of purchase, with no pre-checked boxes
• Retained records — the exact consent language, timestamps, and source — kept at least four years, with seven recommended by defense counsel

This is precisely why My AI Call Center reviews list source and consent records before any campaign launches, and flags or declines bought lists without clear permission records. Telling you plainly that a list won't support the campaign — before you spend anything — is cheaper than learning it from a class-action filing, a category of litigation up 95% year over year with aggregate verdicts exceeding $925 million, per industry tracking.

The Pre-Launch Checklist: Disclosure, Windows, and Records

The difference between a compliant AI calling campaign and a lawsuit often comes down to what happens before the first dial. Here is the checklist that keeps campaigns on the right side of the rules.

1. Verify list source and consent records before dialing. Under the FCC's February 2024 ruling, AI-generated voices count as artificial voices under the TCPA, so every AI call requires prior express consent — and marketing calls require prior express written consent. Critically, consent documentation guidance makes clear that the business making the calls, not the lead seller, bears liability if consent is invalid. An existing business relationship does not exempt AI calls, and co-registration lists carry no legal weight without re-consent.

2. Retain consent documentation for at least four years. The TCPA carries a four-year statute of limitations, and defense counsel recommends seven years. Store the exact consent language shown, timestamps, source URLs, and proof of human interaction — bot-submitted leads carry no valid consent and should be filtered out.

3. Disclose identity at the start of every call. Callers must state their identity and telephone number at the beginning of each call, per legal analysis of the FCC ruling. AI disclosure is not yet federally mandatory, but state laws fill the gap: Texas requires AI disclosure within 30 seconds under SB 140, and California, Florida, Colorado, Illinois, and Utah have their own variants. The safest practice is disclosing AI use on every call, everywhere.

4. Offer opt-out within two seconds. For telemarketing calls, an opt-out mechanism must be available within two seconds of the required disclosures. Revocation requests must then be honored in any reasonable manner — "stop," "quit," "revoke," "unsubscribe" — within 10 business days under rules effective April 11, 2025, according to TCPA compliance guidance.

5. Call only inside approved windows. Residential consumers may only be contacted between 8 a.m. and 9 p.m. in the recipient's time zone, and some platforms enforce an even stricter 8 a.m.–8 p.m. window. State quiet hours and day restrictions layer on top.

A condensed pre-launch checklist:

• List source documented, with consent records matched to every contact
• Consent tier verified — written consent for anything with a marketing purpose
• Identity and AI disclosure scripted into the first 30 seconds
• Opt-out available within two seconds; revocations honored within 10 business days
• Calling windows locked to 8 a.m.–9 p.m. recipient time

The stakes justify the rigor: TCPA damages run $500–$1,500 per call with no aggregate cap, and industry reporting notes class-action filings up 95% year over year.

This is why My AI Call Center builds list and consent review into every campaign before launch — checking list source, consent records, and calling windows, and flagging or declining bought lists without clear permission records. Because the caller bears liability regardless of which vendor dials, every organization should also obtain independent legal guidance before launch, as communications law specialists recommend.

How a Managed Campaign Handles Consent Review for You

Consent review is where most AI calling campaigns quietly fail — and where a managed process earns its keep. The TCPA attaches liability to the business making the call, not the lead seller or the platform, and damages run $500 to $1,500 per call with no aggregate cap. That makes the pre-launch review step non-negotiable.

My AI Call Center builds this review into every campaign before a single call goes out. The list source, consent records, and calling windows are checked against the campaign's purpose, because consent tiers differ: informational calls like appointment reminders need prior express consent, while anything with a marketing or upsell purpose requires prior express written consent that names the business and phone number.

Bought lists rarely survive this review. A TCPA compliance analysis calls co-registration and "warm cold" lists legally meaningless — courts are skeptical of "and our partners" consent language, and the recommended defensible practice is re-consenting every co-reg lead before AI dialing. Lists without clear permission records are flagged and, in most cases, declined. The honest answer comes before you spend anything: "We tell you plainly if the list will not support the campaign."

The review covers more than the list itself:

• Script and disclosure approval — identity, callback number, and opt-out handling are approved by the client before launch; nothing runs until sign-off.
• Opt-out handling — keyword opt-outs are logged and honored immediately, and regulations effective April 2025 require honoring revocation requests within 10 business days.
• Calling windows — calls run only in approved hours in the recipient's time zone, since residential contacts may legally be called only between 8 a.m. and 9 p.m.
• Outcome reporting — dispositioned lists with confirmed, qualified, renewed, opted-out, and no-answer codes, plus opt-out and DNC logs you can retain.

Documentation matters after launch, not just before it. The TCPA carries a four-year statute of limitations, and consent documentation guidance recommends retaining records — the exact consent language shown, timestamps, and source — for at least four years, with defense counsel suggesting seven. A dispositioned outcome report gives you that paper trail.

One caution: no managed process replaces legal advice. Law firm guidance is consistent — TCPA requirements are detailed and use-case specific, and campaign requirements vary by location, industry, contact type, and consent status. Clients should obtain legal counsel for their own situation before launch. A managed campaign handles the operational discipline; the legal sign-off stays yours.