
What qualifies as an autodialer?
Key Facts
- The Supreme Court's unanimous 2021 Facebook v. Duguid ruling defined an autodialer as a device using a random or sequential number generator, narrowing the federal ATDS definition dramatically.
- TCPA violations carry statutory damages of $500 to $1,500 per call, with daily volumes pushing exposure into the millions.
- Federal law bans calls before 8 a.m. or after 9 p.m. local time, and more than 20 states impose even stricter quiet hours, according to TCPA compliance guidance.
- The FCC requires honoring revocation requests within 10 business days, with proposals to shorten that window to 7 days, per recent consent revocation rules.
- AI-generated voices are treated as artificial voices under the TCPA, requiring prior express consent even when the dialer is not an ATDS, according to FCC mandates.
- Standard predictive dialers and stored-list systems generally fall outside the federal ATDS definition after Duguid, as Kelley Drye's legal analysis explains.
- State 'mini TCPAs' can define autodialers more strictly than federal law, so multi-state campaigns require per-state review, according to compliance experts.
Why the Autodialer Question Still Trips Up Callers
A single autodialer claim can cost you $500 to $1,500 per call, and at daily volume that exposure climbs into the millions fast. The trouble is that most teams never check which definition actually applies to their dialing, so they either over-restrict legitimate outreach or assume consent rules don't reach them.
The controlling federal standard comes from the Supreme Court's unanimous 2021 decision in Facebook v. Duguid. Under that ruling, a device qualifies as an autodialer only if it can store or produce phone numbers using a random or sequential number generator. That resolved a long circuit split, and the Court sided with the narrow reading.
Practical effect: systems that simply dial from a stored list generally fall outside the ATDS definition. The Court reasoned that a broad reading would sweep in virtually every modern cellphone, which Congress never intended, as Kelley Drye's analysis of the ruling explains. Predictive dialers and list-based platforms sit in different territory than they did before 2021.
Here's the part that catches people: Duguid narrowed one definition and left another untouched. The TCPA separately bars calls using an artificial or prerecorded voice, and AI-generated voices count as artificial. So a campaign can avoid the ATDS label entirely and still need prior express consent. That distinction drives how My AI Call Center scopes campaigns, verifying list source and consent records before launch rather than assuming a non-ATDS system clears the bar.
A few points worth keeping straight:
- Federal calling hours run 8 a.m. to 9 p.m. local time, and more than 20 states impose tighter windows.
- Revocation requests must be honored within 10 business days, with the FCC proposing to shorten that to 7.
- State "mini TCPAs" can define autodialers more strictly than federal law, so multi-state campaigns need per-state review.
The honest answer to "what qualifies as an autodialer" is that the federal definition is narrower than most teams assume, while the consent obligations around AI voices remain broad. Treat those as two separate questions, and the compliance picture gets clearer.
The Duguid Standard: What the Supreme Court Actually Said
For years, businesses couldn't agree on what an "autodialer" even was. The Supreme Court settled it on April 1, 2021, in Facebook, Inc. v. Duguid, a unanimous decision that narrowed the definition dramatically and changed how every outbound calling operation should think about compliance.
The ruling came down to grammar. The TCPA defines an automatic telephone dialing system as equipment that can store or produce telephone numbers "using a random or sequential number generator." Facebook argued the phrase modified only "produce." The Court, in an opinion by Justice Sotomayor, applied the series-qualifier canon and held the phrase modifies both verbs. A device qualifies as an ATDS only if it can store or produce numbers using a random or sequential number generator. As Kelley Drye's analysis of the ruling explains, the Court stuck to the plain grammar of the statute itself.
The stakes were high because courts had split for years. The Second, Sixth, and Ninth Circuits read the statute broadly, treating any device that stored and dialed numbers as an autodialer. The Third, Seventh, and Eleventh Circuits required a random or sequential generator. The Supreme Court adopted the narrow view, and more than 20 parties filed amicus briefs hoping to influence the outcome.
Why does the narrow reading matter? The Court reasoned that Congress targeted a specific problem: equipment that risks dialing emergency lines randomly or tying up sequentially numbered lines at a single entity. A broad definition would sweep in virtually every modern cellphone, which the Court said would "take a chainsaw to these nuanced problems when Congress meant to use a scalpel." That concern echoed the D.C. Circuit's earlier warning in ACA International v. FCC (2018) that an expansive reading would be an "eye-popping sweep."
The practical result for businesses running outbound campaigns:
- Standard predictive dialers generally don't qualify as ATDS unless they can generate numbers randomly or sequentially.
- Systems that dial from a stored contact list, without number generation capacity, fall outside the federal ATDS definition.
- The FCC's earlier position that dialing stored numbers without human intervention counted as an autodialer no longer holds under the federal standard.
- The ruling resolved the circuit split, so the same system can't be an autodialer in one state and not another under federal law.
Here's the part that trips people up: narrowing the ATDS definition didn't narrow the TCPA itself. The statute separately prohibits calls made with an artificial or prerecorded voice, and that prohibition stands regardless of how the numbers were dialed. As Wiley's alert on the decision notes, a system can dodge the ATDS label and still face TCPA liability for artificial voice calls made without prior express consent. AI-generated voices are treated as artificial voices, so consent requirements apply even when the dialing technology itself is not an autodialer.
That's why list discipline matters as much as dialing technology. My AI Call Center runs structured campaigns against approved, permissioned, or reviewed lists only, with consent records checked before any campaign launches, because the Duguid standard is a floor, not a shield. Violations still carry statutory damages of $500 to $1,500 per call, and state "mini TCPAs" can impose stricter definitions than the federal one. Clear on the dialer, careful with the voice, strict on the list: that's the workable path through this standard.
The Catch: AI Voices Are Regulated Separately
The Catch: AI Voices Are Regulated Separately
You might think the Duguid ruling gives you a free pass for AI calls. It does not. The Supreme Court narrowed the definition of an autodialer, but it left the TCPA’s rules on artificial voices completely intact. If your system uses an AI-generated voice, you still need prior express consent. This distinction is the biggest compliance trap for modern outbound campaigns.
The law treats AI voices as "artificial" under federal regulations. Even if your dialing software doesn’t use a random or sequential generator, the voice itself triggers strict consent requirements. Legal analyses confirm that the Duguid decision only addressed the dialing mechanism, not the content of the call. Ignoring this separation exposes you to significant financial risk.
Violations carry statutory damages of $500 to $1,500 per call. For high-volume campaigns, this exposure can quickly reach into the millions. Recent legal reviews highlight that these penalties apply regardless of whether the device qualifies as an ATDS. The voice is the deciding factor, not the dialer.
State laws add another layer of complexity. Many states have enacted "mini TCPAs" with stricter rules than the federal baseline. These local laws often impose tighter calling hours and specific registration requirements. You must check the regulations in every state where your contacts reside.
Consider these specific regulatory constraints when planning your campaigns:
- Federal law bans calls before 8 a.m. or after 9 p.m. local time.
- More than 20 states enforce even stricter quiet hours than the federal standard.
- The FCC requires honoring revocation requests within 10 business days.
- Automated systems must recognize standardized opt-out words like "stop" and "cancel."
The FCC has also updated rules on how consent revocation works. Callers can now designate exclusive opt-out methods, but they must still honor requests quickly. The current window is 10 business days, though proposals to shorten this to 7 days are under review. New FCC orders mandate that automated systems, including AI voices, have robust opt-out mechanisms. If a recipient says "stop," the system must log it and honor the request immediately.
At My AI Call Center, we treat AI-generated voices as artificial voices from day one. We verify consent records before any campaign launches, ensuring your list supports the specific technology you are using. We check state-specific quiet hours and registration rules for every contact in your database. This approach keeps your campaigns compliant, even when federal definitions shift.
Compliance isn't just about avoiding fines; it's about protecting your brand and your data. By respecting the separate rules for AI voices, you build trust with your audience. You ensure that every call is welcome, not a nuisance. This discipline is the foundation of a sustainable outbound strategy.
How to Run Compliant Outbound Campaigns in Practice
Compliance isn't a checkbox you clear once. It's a set of operational habits you repeat on every campaign. Here's how that looks in practice.
Before any campaign launches, verify the list source and consent records. If a list was bought and comes without clear permission records, flag it and decline it. That stance matters because TCPA violations carry statutory damages of $500 to $1,500 per call, and daily volume can push total exposure well into the millions, according to a TCPA compliance guide. No list, no launch.
Then check your calling windows. Federal law bars calls before 8 am or after 9 pm local time, and more than 20 states impose tighter restrictions. My AI Call Center honors state-specific quiet hours and windows rather than defaulting to the federal floor.
On every call, disclose that the voice is AI-assisted and give recipients a clear way out. The FCC has mandated automated opt-out mechanisms for artificial and prerecorded voice calls, including AI-generated ones, per the FCC's revised consent revocation rules. Keyword opt-outs like STOP and REVOKE get logged immediately and honored across all campaigns.
A practical pre-launch sequence looks like this:
- Confirm list source and consent records before anything else
- Decline bought lists without clear permission documentation
- Map state-specific quiet hours and calling windows
- Approve AI disclosure and keyword opt-out handling in the script
- Route opt-outs and DNC requests into client records immediately
Keep your legal counsel in the loop. Requirements vary by state, industry, contact type, and consent status, and state "mini TCPAs" can impose stricter definitions than the federal standard. Your counsel, not a vendor, owns that call.
Frequently Asked Questions
What actually counts as an autodialer under federal law?
Does my predictive dialer count as an autodialer?
If my dialer isn't an autodialer, do I still need consent for AI voice calls?
How much can I be fined for a TCPA violation?
What are the legal calling hours for outbound campaigns?
How quickly do I have to honor someone who says 'stop' or asks to opt out?
The Narrow Dialer and the Wide Voice: Two Questions, One Discipline
The federal answer to "what qualifies as an autodialer" is narrower than most teams assume: under Facebook v. Duguid, a device is an ATDS only if it can store or produce numbers using a random or sequential number generator, which leaves most stored-list and predictive dialers outside the definition. But that clarity comes with a catch, the TCPA's separate ban on artificial or prerecorded voices is untouched, and AI-generated voices count as artificial, so prior express consent still governs the call itself. Add state mini-TCPAs, quiet hours, and opt-out windows, and the workable path is the same one My AI Call Center builds its campaigns around: clear on the dialer, careful with the voice, strict on the list. Before your next launch, treat list source, consent records, calling windows, and opt-out handling as separate checks, not one blanket assumption. If you want that review done with you rather than guessed at, start with a free campaign review at myaicallcenter.app, where the whole number is known before you approve anything.