The Short Answer: Restricted, Not Banned

Automated phone calls are not banned in the United States. They are heavily restricted. The Telephone Consumer Protection Act, enacted in 1991, regulates specific technologies rather than prohibiting all automated outreach. Your legality depends on five variables: the technology used, the number called, the message content, the consent obtained, and any applicable exemptions.

This distinction matters because of the financial stakes involved. Statutory damages run $500 per violation, which can triple to $1,500 for willful or knowing violations. A single non-compliant campaign of 10,000 calls creates potential exposure between $5 million and $15 million. This math explains why compliance is a primary operational concern, not a secondary legal formality.

The regulatory landscape shifted significantly in February 2024. The Federal Communications Commission issued a Declaratory Ruling confirming that AI-generated voices count as "artificial voices" under the TCPA. This means AI voices trigger mandatory consent requirements regardless of whether an autodialer is used. The statute does not allow carve-outs for technologies that provide the equivalent of a live agent. Any AI-generated voice automatically triggers these consent rules, no matter how human-like it sounds.

Understanding the five variables helps clarify where the legal line is drawn.

• Technology: Determines if the call uses an autodialer or artificial voice, both of which face strict consent hurdles.
• Number Called: Wireless numbers and residential lines often require prior express consent, while business lines have different protections.
• Consent: The dividing line between legal and illegal calls, with written consent often required for marketing purposes.
• Exemptions: Specific categories like healthcare reminders or debt collection may have different rules than general telemarketing.

My AI Call Center operates within this framework by running structured campaigns only against approved, permissioned, or reviewed lists. We check list source and consent records before any campaign launches to ensure the technology matches the consent tier. This approach aligns with the current legal reality where consent is the primary defense.

The penalty structure drives the need for precision. Courts view each unconsented call as a separate violation. This makes class actions the dominant risk for operators who rely on indiscriminate cold calling. By focusing on permissioned lists and clear consent records, you mitigate the risk of triggering these per-call penalties. The law does not offer an automatic recovery for every unwanted contact, but it does provide significant remedies for proven violations.

Regulatory clarity continues to evolve, but the core principle remains consistent. Automated calls are legal when you respect the specific consent requirements for the technology and audience involved. The February 2024 ruling closed the door on the idea that AI voices might escape these rules. Now, the focus is on which consent standard applies to which call type. Operators must navigate this by ensuring their consent records are robust, traceable, and specific to the business making the call.

Consent Is the Dividing Line, and It Just Got Complicated

If you run outbound calling campaigns, consent is the rule that decides whether your calls are legal or a lawsuit waiting to happen. And in 2026, the rules around that consent just split in two, depending on where you call.

The TCPA has always used a two-tier consent standard. Autodialed or prerecorded calls to wireless numbers need prior express consent, while advertising and telemarketing calls using those technologies need the higher bar of prior express written consent, defined as a signed agreement that discloses the caller and states signing is not a condition of purchase, according to the TCPA's regulatory framework.

That second tier is now under attack. In February 2026, the Fifth Circuit ruled in Bradford v. Sovereign Pest Control of Texas that the statute's text only requires "prior express consent," finding the FCC overstepped when it created the written-consent requirement in 2012. The practical effect: oral consent may now satisfy the law for marketing calls in Texas, Louisiana, and Mississippi.

Meanwhile, the Eleventh Circuit, in Insurance Marketing Coalition v. FCC, vacated the 2023 one-to-one consent rule but left the written-consent requirement for advertising robocalls fully intact. So the same AI marketing call can be lawful with a verbal "yes" in Houston and a violation in Atlanta.

Here's the practical answer: don't chase the split. Because campaigns cross state lines and the circuits disagree, the compliance-safe posture is to meet the higher written standard everywhere. That's how we approach it at My AI Call Center: consent records and list source are reviewed before any campaign launches, and lists without clear permission documentation are flagged, and in most cases declined.

One trap catches even experienced operators: the established business relationship exemption. An existing business relationship (generally 18 months from the last purchase or 3 months from an inquiry) exempts manual calls from Do-Not-Call restrictions. It does nothing for AI calls. As TCPA counsel note, a live salesperson can call a 16-month-old customer on the DNC list; an AI agent calling that same person without separate consent is a violation.

Consent rules are also changing on the revocation side. An FCC order released October 1, 2026, replaces the "revoke all" rule with category-specific opt-out rules: an opt-out from an informational call (say, an appointment reminder) may be treated as applying to that category only, while marketing opt-outs still revoke all marketing consent, including from affiliated entities. The 10-business-day honor window is unchanged, though a pending FCC proposal could shorten it to seven.

The takeaway is simple. Get written consent, keep the records for years, and treat every "stop" or "revoke" as urgent regardless of category. With penalties running $500 to $1,500 per call, the paperwork is the protection.

Where Calls Actually Go Wrong: The Failure Points Courts Target

Most automated-call lawsuits don't hinge on exotic legal theories. They hinge on paperwork: consent that can't be traced, disclosures that never happened, and STOP requests that got ignored. The settlements tell you exactly where courts are looking.

The vendor chain is the first failure point. QuoteWizard's $19 million settlement has become the reference point for what happens when a company can't trace consent through its lead generators and vendors, according to TCPA litigation analysis. The pattern is now explicit in class filings: Lamb v. Mortgage One Funding proposes a class covering calls made by the company "or from any of the company's vendors, lead generators, or agents." The brand on whose behalf the call was made bears the risk, no matter who actually dialed.

Consent gaps on live transfers come next. A named violation pattern in current litigation involves an AI agent that qualifies a lead, then transfers to a live representative who lacks proper consent for the conversation that follows, as attorney-authored compliance analysis documents. Gen Digital's $9.95 million settlement in January 2026 shows the dollar scale: prerecorded calls to people who were never customers at all.

The other recurring failure points are narrower but just as expensive:

• Missing or late AI disclosure. Texas requires disclosure within 30 seconds; Florida requires written consent that explicitly references AI use.
• Ignored STOP requests. Consumers may revoke consent "in any reasonable manner," and callers must honor it within 10 business days.
• "And our partners" co-registration lists. Courts are skeptical of consent where the consumer couldn't identify the actual buyer at the moment they signed.

B2B callers get no pass. Business-to-business calls and texts face the same TCPA wireless restrictions as consumer calls, and roughly half of states run their own broader mini-TCPA statutes, per compliance guidance. A recruiting firm dialing a candidate's cell with an artificial voice needs the same consent a retailer does.

And there's a quieter consequence: your non-compliant calls may never ring. Since a 2019 FCC ruling, carriers can block unwanted calls by default using "reasonable call analytics," including complaint volume, TCPA-violation patterns, and missing SHAKEN/STIR call signing, per the complete TCPA compliance guide. The call isn't just illegal; it's dead air.

This is why list discipline is a legal control, not a quality preference. My AI Call Center reviews list source and consent records before any campaign launches, and bought lists without clear permission records are flagged and usually declined. Statutory damages run $500 per violation, up to $1,500 when willful, so a single sloppy 10,000-call campaign carries $5M to $15M in exposure. The same audit trail that protects a compliant operator convicts a non-compliant one.

How to Run Compliant AI Calling Campaigns

Consent is your first line of defense, and it has to be specific. Courts are skeptical of consents that name "and our partners" when the consumer can't identify the buyer at the moment of consent, so every list should carry written consent that names your business directly. Requiring prior express written consent on every list, even where circuits like the Fifth Circuit's Bradford ruling permit oral consent, removes the need to adjudicate jurisdiction contact by contact.

Disclosure comes next, and it belongs in the script by default rather than as a state-by-state toggle. Texas requires AI disclosure within 30 seconds, and California, Florida, Colorado, Illinois, and Utah each have their own variants, while a pending FCC rulemaking proposes mandatory disclosure at the start of every AI-generated call. Standardizing disclosure inside that 30-second window satisfies the strictest state threshold and pre-positions your campaigns for the federal mandate.

Revocation handling should be mapped before launch, not improvised mid-campaign. The FCC's October 2026 Updated Order permits category-specific opt-outs for informational messages while marketing opt-outs still revoke all marketing consent, so each campaign type (payment reminders, appointment reminders, renewal calls, surveys) needs a defined message category with revocations logged against it. Build toward a seven-business-day processing target, since a pending proposal could shorten the current 10-business-day window.

• Carry consent records through every live transfer, because consent gaps on transfers are a named violation pattern and the $19 million QuoteWizard settlement shows where vendor-chain failures land.
• Retain consent and call records for seven years, even though the TCPA statute of limitations runs four, because that audit trail is what converts a defense theory into a defensible position.
• Treat your dispositioned lists, opt-out logs, and DNC records as legal instruments rather than reporting extras.

For teams running managed campaigns on approved, permissioned lists, My AI Call Center builds these controls into the process before launch: list source and consent records reviewed, disclosure scripted, opt-outs logged and honored across campaigns. The campaigns start at 9¢ per connected minute, and the full number is quoted before you approve anything.

What a Compliant Campaign Looks Like in Practice

Knowing the rules is one thing. Building a campaign that actually follows them, call after call, is where most operators stumble. The good news: compliance is mostly a process problem, and process can be checked before a single call goes out.

At My AI Call Center, every campaign starts with the list. Before launch, we review the list source and the consent records behind it. That matters because courts have grown skeptical of consents naming "and our partners" where the consumer can't identify the actual buyer at the moment of consent. Bought lists without clear permission records get flagged, and in most cases declined. We tell you plainly if the list won't support the campaign, before you spend anything.

The stakes justify that discipline. Statutory damages run $500 per violation, trebled up to $1,500 for willful violations, and a non-compliant 10,000-call campaign carries $5M to $15M in potential exposure according to attorney analysis of AI voice compliance. Class-action settlements in 2025 and 2026 have landed in the $5M to $20M range.

Once the list clears review, the operational guardrails lock in:

• Calls run only inside approved windows, 8 a.m. to 9 p.m. in the recipient's local time, with state-specific quiet hours honored.
• Every call includes AI disclosure, and recipients can ask if the call is AI-assisted, request a human, or opt out at any point.
• Keyword opt-outs like STOP and REVOKE are honored immediately, and DNC requests carry across all campaigns into your DNC records.
• Opt-out and DNC logs ship as standard campaign deliverables, alongside the dispositioned contact list and outcome counts.

That last point deserves emphasis. As defense counsel put it, "the same audit trail that protects a compliant operator convicts a non-compliant." The TCPA statute of limitations is four years, but defense attorneys recommend keeping consent records for seven. Your campaign reports aren't just performance data; they're your evidence if a dispute ever arrives.

One honest caveat: requirements vary by location, industry, contact type, and consent status, and the rules keep moving. A 2026 circuit split means courts even disagree on whether marketing calls require written consent. We meet the higher written standard everywhere, but your situation may have specifics no service provider can resolve for you. Get appropriate legal guidance before launch.

If you want a second set of eyes on your list and consent posture, the first campaign review is free. We'll tell you what the rules require for your specific case and what the whole campaign costs before you approve anything.