TCPA and Telemarketing Rules · September 28, 2026 · GrowthPros

Is outbound AI calling illegal?

Is AI calling illegal? Not if you have consent. Learn FCC TCPA rules for AI voice calls, common compliance mistakes, and how to call leads legally withi...

Flat illustration of a smartphone with a call icon, shield and consent document motifs in lime and olive green, headlined CALL LEGALLY.

Key Facts

  • TCPA penalties range from $500 to $1,500 per violating call with no aggregate cap
  • TCPA class-action filings increased 95% year over year
  • Aggregate TCPA verdicts have exceeded $925 million across the docket
  • Gen Digital settled for $9.95 million in January 2026 for TCPA violations
  • QuoteWizard paid $19 million in a TCPA settlement
  • Defense counsel recommends retaining consent records for up to seven years
  • AI voice marketing calls require prior express written consent in 47 states

The Fear Stopping Businesses From Using AI Calling

If you run a dealership, an insurance agency, or a contracting business, you've probably seen the headline: the FCC made AI robocalls illegal. For a lot of business owners, that headline settles the matter — AI voice follow-up is off-limits, and the leads sitting in your CRM stay untouched. That conclusion feels safe. It's also wrong.

To be fair, the fear is rational. In February 2024, the FCC unanimously ruled that AI-generated voices count as "artificial or prerecorded voice" under the Telephone Consumer Protection Act. And TCPA penalties are not gentle: compliance analysis puts the cost at $500 to $1,500 per call, with no aggregate cap. One bad campaign can snowball into a seven-figure liability.

The litigation numbers back that up. According to recent TCPA filings data, class-action lawsuits are up 95% year over year, and aggregate verdicts across the docket have exceeded $925 million. Recent settlements tell the same story — QuoteWizard paid $19 million, and Gen Digital settled for $9.95 million in January 2026.

The stakes, in short:

  • $500–$1,500 per violating call, with no cap on total damages
  • TCPA class-action filings up 95% year over year
  • A four-year statute of limitations, with defense counsel recommending seven years of consent records

But here's what the headlines leave out. As legal analysis of the ruling makes clear, the FCC did not create a new prohibition — it clarified that AI voices fall under the same existing robocall rules that have governed automated calls for decades. AI calling to a consented lead is not illegal; AI calling to someone who never opted in is. The technology isn't the violation. Missing consent is.

That distinction matters enormously for how you buy and work leads. A lead with a documented consent trail — disclosure text, timestamp, IP address, and the named contacting party — can be called by an AI voice within minutes of submission, legally. A "warm" co-registration list with no consent records cannot, no matter how promising the vendor sounds. Compliance lives in the consent record, not in the voice on the call.

This is why GrowthPros attaches a full consent trail to every lead before any AI follow-up happens, and scrubs lists against the DNC before a single dial goes out. The businesses getting burned aren't the ones using AI — they're the ones skipping the consent documentation entirely and hoping nobody notices. The rest of this article shows exactly where the legal lines sit, and how to stay on the right side of them.

What the Law Actually Says: AI Voices Are Regulated, Not Banned

Outbound AI calling isn't illegal — but treating it like it's unregulated is how companies end up writing seven-figure settlement checks. The FCC settled the core question on February 8, 2024, with a Declaratory Ruling that classified AI-generated voices as "artificial or prerecorded voice" under the TCPA. As Elias Law Group notes, the ruling didn't create a new prohibition; it simply confirmed AI calls fall under the same consent rules as traditional robocalls.

What that means in practice: AI voice marketing calls to wireless numbers require prior express written consent in 47 states, according to a TCPA compliance analysis. The Fifth Circuit's Bradford v. Sovereign Pest Control decision carved out an exception holding that oral consent suffices in Texas, Louisiana, and Mississippi — but no operator should build a national program on the narrowest reading of the law.

The most expensive trap is the Established Business Relationship (EBR) exemption. A live agent may legally call a past customer under EBR, but an AI agent needs separate consent for the exact same contact — a distinction compliance practitioners call the single most expensive misunderstanding in AI outbound. Reactivating a dormant customer list with AI voice without re-consent is exactly the scenario the TCPA's $500–$1,500 per-violation penalties were built for.

States layer additional requirements on top:

  • Texas SB 140 (effective September 2024) requires AI disclosure within the first 30 seconds of a call.
  • Texas TRAIGA (HB 149), effective January 1, 2026, adds further AI transparency obligations.
  • A federal in-call AI disclosure rule is expected within 12–24 months, per a pending FCC rulemaking.

One rule got simpler: the FCC's "one-to-one consent" requirement was vacated by the Eleventh Circuit in January 2025 and never took effect. But as compliance infrastructure experts point out, state mini-TCPA laws keep filling federal gaps, so the volatility isn't over.

The takeaway for buyers of leads and calling programs: legality lives in the consent trail, not the dialer. That's why GrowthPros attaches a consent record — disclosure text, timestamp, IP address, and named contacting party — to every lead before any AI voice, SMS, or email touches it, and why reactivation campaigns target only opted-in relationships, never cold lists. Compliance is a data problem before it's a legal one, and the FCC's own ruling documents make clear the agency expects provable consent, not assumptions.

Where Businesses Get Sued: The Expensive Compliance Mistakes

Most TCPA lawsuits don't come from businesses that ignored the law on purpose. They come from businesses that assumed consent existed somewhere upstream — in a co-registration form, an old CRM list, a vendor's "compliant" dialer. That assumption costs $500 to $1,500 per call, with no aggregate cap and a four-year statute of limitations, which is why TCPA class-action filings are up 95% year over year and aggregate verdicts have exceeded $925 million.

The first expensive mistake is trusting "warm cold" lists. Co-registration leads — someone clicked a checkbox buried in a quiz or sweepstakes form — have no legal standing for AI voice calls. Legal experts across the industry agree there is no defensible shortcut here: re-consent before AI dialing is the only approach that holds up in court.

The second mistake is assuming an Established Business Relationship (EBR) covers AI outreach. It doesn't. A live agent may call a past customer under EBR, but an AI agent needs separate consent for the same contact — widely described as the single most expensive misunderstanding in the AI outbound playbook. If you're reviving a dormant list with AI voice, that old opt-in from 2022 is not enough.

The third mistake is believing you can outsource liability. The Lamb v. Mortgage One Funding case establishes that entities commissioning AI calls bear liability regardless of whether they or their vendors placed the calls. Compliance cannot be delegated to a dialer platform.

Finally, even fully consented campaigns can violate the rules on mechanics alone:

  • 3% abandonment cap — no more than 3% of answered calls may be abandoned, measured over a 30-day FCC campaign window or per day under the FTC's Telemarketing Sales Rule.
  • 2-second connect rule — a call is "abandoned" if it connects but delivers no response within two seconds of the consumer's greeting.
  • DNC scrubbing must happen at call initiation in real time, not in a nightly batch.

As Teams Plus President Ricardo J. Ordonez puts it, compliance at scale requires call logging, consent linkage, real-time DNC scrubbing, and abandonment tracking built into the infrastructure — not a policy document. That's the pattern in every recent settlement, from QuoteWizard's $19M to Gen Digital's $9.95M: the consent infrastructure wasn't there when plaintiffs' counsel came asking.

This is why GrowthPros attaches a consent record — disclosure text, timestamp, IP address, and named contacting party — to every lead it delivers, and runs dead-lead reactivation only against pre-existing, opted-in relationships. Consent built in, not bolted on, is the difference between a growth channel and a class action.

How Compliant AI Calling Works in Practice: The GrowthPros Playbook

Knowing the rules is one thing; running a lead-gen pipeline that survives a TCPA class action is another. With TCPA filings up 95% year over year and penalties of $500–$1,500 per violation with no aggregate cap, compliance can't live in a policy document — it has to be built into the infrastructure itself, as carrier-grade voice providers point out.

That's exactly how GrowthPros approaches outbound AI calling. Every lead carries a complete consent record — the disclosure text the consumer saw, the timestamp, the IP address, and the named contacting party. If a call is ever challenged, the defense isn't a verbal assurance; it's the documented trail attached to the lead when it lands in the client's CRM. Defense counsel recommends retaining these records for up to seven years, which only works if consent linkage happens at capture time, not reconstructed later.

Before any dial happens, lists are scrubbed against the DNC in real time rather than in overnight batches. This matters because the FCC's February 2024 Declaratory Ruling confirmed AI-generated voices fall squarely under TCPA robocall rules — there's no technicality to hide behind.

Opt-outs are treated as permanent and channel-wide. When a consumer says "stop" — by text, voice, or any reasonable means — the revocation is honored within the 10-business-day window guidance and applied across SMS, voice, and email simultaneously, with no re-contact from any channel afterward.

The reactivation side follows the same discipline, with one bright line:

  • Only pre-existing, opted-in relationships are contacted — the client's own dormant CRM list, never a purchased cold file
  • "Warm" co-registration lists have no legal standing for AI dialing, so they're rejected outright
  • Established Business Relationship exemptions don't carry over to AI voice — a live agent may call a past customer, but an AI agent needs its own consent basis

The practical payoff: a compliant pipeline and a fast one aren't in conflict. Consent-recorded leads still get AI voice, SMS, and email follow-up inside the five-minute speed-to-lead window, 24/7 — and typically 8–15% of a dormant opted-in database re-engages without a single legally exposed call. The promise isn't that any lead closes; it's that every lead is qualified, documented, and followed up inside the promised window, with its consent trail intact.

The legal question has an answer, and the commercial question has a number. Outbound AI calling is lawful when built on documented consent — and the payoff for doing it right is enormous.

The math on speed is hard to ignore. Contacting a lead within five minutes makes contact roughly 100x more likely than waiting thirty, and about 78% of buyers choose whoever responds first. Meanwhile, TCPA penalties run $500–$1,500 per violation with no aggregate cap, and class-action filings are up 95% year over year. The businesses that win are the ones that treat speed and compliance as the same system, not competing priorities.

A defensible AI speed-to-lead operation comes down to a few non-negotiables:

  • Consent records attached to every lead — disclosure text, timestamp, IP address, and named contacting party, retained for the long haul.
  • Real-time DNC scrubbing and consent linkage built into the calling infrastructure, not a policy document.
  • Opt-outs honored immediately and permanently across SMS, voice, and email.
  • No reliance on "established business relationship" for AI calls — a live agent may call a past customer, but an AI agent needs separate consent.

As compliance practitioners put it, this has to live in the infrastructure. Batch processes and good intentions don't survive a four-year statute of limitations and a plaintiff's attorney with a subpoena.

That's exactly how GrowthPros approaches it. Every lead is consent-recorded and DNC-scrubbed before delivery, then followed up by AI voice, SMS, and email inside the five-minute window — because the fastest legal response is usually the winning one. And if you're sitting on a dormant, opted-in CRM list, reactivation is typically the cheapest qualified leads you'll ever generate.

The next step is simple: book the 15-minute qualification call. We'll review your niche, your consent posture, and whether your existing database is a reactivation candidate worth running. It's free, honest about fit, and commits you to nothing.

Exclusive leads by niche, followed up in minutes — including the leads you already paid for. Reach us at [email protected] or explore more at growthpros.marketing/insights.

Frequently Asked Questions

Did the FCC ban AI robocalls in 2024?
No. The FCC's February 8, 2024 Declaratory Ruling didn't create a new ban — it clarified that AI-generated voices count as "artificial or prerecorded voice" under the TCPA, so AI calls follow the same consent rules that have governed robocalls for decades. AI calling to a consented lead is legal; calling someone who never opted in is what gets you sued. Elias Law Group's analysis of the ruling makes this distinction explicit.
What consent do I need before making AI voice calls to leads?
For marketing calls to wireless numbers, you need prior express written consent in 47 states, according to a TCPA compliance analysis. The Fifth Circuit's Bradford v. Sovereign Pest Control decision allows oral consent in Texas, Louisiana, and Mississippi — but no operator should build a national program on the narrowest reading of the law. A documented consent trail (disclosure text, timestamp, IP address, and named contacting party) is what makes a lead legally callable by AI voice.
How much can a TCPA violation actually cost my business?
Between $500 and $1,500 per violating call, with no aggregate cap — and a four-year statute of limitations. Recent TCPA filings data shows class-action lawsuits up 95% year over year with aggregate verdicts exceeding $925 million, including settlements like QuoteWizard's $19 million. One bad campaign can snowball into seven-figure liability, which is why GrowthPros attaches a full consent record to every lead before any AI follow-up happens.
Can I use AI voice to call my past customers without new consent?
No — this is widely described as the single most expensive misunderstanding in AI outbound. A live agent may legally call a past customer under the Established Business Relationship (EBR) exemption, but an AI agent needs separate consent for the exact same contact. Reactivating a dormant customer list with AI voice without re-consent is exactly the scenario the TCPA's $500–$1,500 per-violation penalties were built for.
If my vendor or dialer platform places the calls, am I still liable?
Yes. The Lamb v. Mortgage One Funding case establishes that entities commissioning AI calls bear liability regardless of whether they or their vendors placed the calls — compliance cannot be delegated to a dialer platform. Even fully consented campaigns must follow call mechanics: no more than 3% of answered calls abandoned (measured over a 30-day FCC window), a 2-second connect rule, and real-time DNC scrubbing built into the infrastructure, not nightly batches.
Are "warm" co-registration lead lists okay for AI calling if the vendor says they're compliant?
No. Co-registration leads — someone clicked a checkbox buried in a quiz or sweepstakes form — have no legal standing for AI voice calls, and legal experts agree there's no defensible shortcut. Compliance practitioners are unanimous: re-consent before AI dialing is the only approach that holds up in court. That's why GrowthPros rejects warm co-registration lists outright and runs reactivation only against pre-existing, opted-in relationships.

The Verdict: AI Calling Is Legal — Sloppy Consent Is What Gets You Sued

So, is outbound AI calling illegal? No — but the question was never really about the technology. The FCC's 2024 ruling simply confirmed that AI voices follow the same consent rules robocalls have faced for decades. What gets businesses sued is missing consent: co-registration lists with no paper trail, dormant CRM lists revived without re-consent, and the mistaken belief that an established business relationship covers AI outreach when it doesn't. With penalties of $500–$1,500 per violating call, no aggregate cap, and class-action filings up 95% year over year, compliance has to live in your infrastructure — real-time DNC scrubbing, consent records attached to every lead, and opt-outs honored across every channel. The good news: a legally defensible pipeline is also a fast one. GrowthPros builds that documentation into every lead before any AI voice, SMS, or email goes out, so speed-to-lead and TCPA safety become the same system. Your next step is simple: audit your current consent posture, then book the free 15-minute qualification call to see whether your niche — or your dormant database — is a fit. It commits you to nothing.

This article is general information, not legal or financial advice. Benchmark figures are directional industry data, not guarantees of results.

Start

More booked calls. Not more form fills.

Tell us your niche and your goal. We will show you realistic volume, exclusivity options, and what follow-up looks like on a live call — no pressure, no 40-page deck.