The Legal Definition: When a Call Counts as Automated

If your AI agent can crack jokes, answer questions, and pause naturally like a coworker, it must not be a robocall, right? Legally, that assumption is wrong — and it's one of the most expensive misconceptions in outbound calling today.

Under the Telephone Consumer Protection Act (TCPA), a call counts as "automated" if it uses an automatic telephone dialing system (ATDS) or contains an artificial or prerecorded voice. That definition matters because it triggers strict consent requirements, calling-window limits, and disclosure obligations. The technology's sophistication is irrelevant to the legal test — what matters is how the call was placed and what kind of voice delivered it.

The FCC removed any remaining ambiguity in its February 2024 Declaratory Ruling, which explicitly classified AI-generated voices as artificial or prerecorded voices under the TCPA. As compliance analysis of the ruling notes, the FCC stated the statute "does not allow for any carve out of technologies that purport to provide the equivalent of a live agent." In other words, a conversational AI that responds in real time is treated the same as a canned robocall from 2005.

The stakes are substantial. TCPA statutory damages run $500 to $1,500 per call with no aggregate cap, and FCC fines can reach $23,727 per violation, according to a TCPA compliance guide for AI voice calls. A single non-compliant campaign of 100,000 calls could theoretically expose a business to $50 million or more in damages. TCPA filings have also surged, with industry trackers reporting a 95% year-over-year increase.

Once a call is classified as automated, several obligations attach:

• Consent tiers by purpose — marketing AI calls require Prior Express Written Consent in 47 states, while informational calls need Prior Express Consent, per legal compliance guidance.
• Calling windows limited to 8:00 AM–9:00 PM in the recipient's local time zone.
• Opt-out mechanisms offered within two seconds of company identification, with revocations honored within 10 business days (TCPA documentation).
• Consent records retained for at least five years, with some states requiring up to seven.

One further trap catches many businesses: an established business relationship does not exempt AI calls. A live agent may call a past customer on the DNC list, but an AI agent cannot without separate consent — and liability extends to the company on whose behalf calls are made, even when a vendor places them.

This is why disciplined providers treat AI calls as fully regulated from day one. My AI Call Center, for example, checks list source and consent records before any campaign launches, declines bought lists without clear permission records, and builds AI disclosure and keyword opt-outs into every script. Managed outbound calling campaigns against approved, permissioned lists start at 9¢ per connected minute — plan your campaign with the goal, list, and consent questions answered up front. The lesson for any organization: assume your AI voice is an artificial voice, because the FCC already does.

What Automation Legally Triggers: Consent, Disclosure, and Opt-Outs

When a call is classified as automated, it triggers a specific set of legal obligations designed to protect consumers from unwanted or deceptive communications. Under the TCPA, any call using an artificial or prerecorded voice—including AI-generated voices—is considered an automated call and requires prior express consent, with requirements varying by call purpose and jurisdiction.

For marketing calls, prior express written consent (PEWC) is required in 47 states, while informational or transactional calls may proceed with prior express consent (PEC) according to compliance experts. Notably, an established business relationship does not exempt AI calls from these consent rules, meaning even past customers must provide separate consent for AI-initiated outreach as clarified by regulatory guidance.

Automated calls are restricted to calling windows of 8:00 AM to 9:00 PM in the recipient’s local time zone, and opt-out mechanisms must be offered within two seconds of company identification per technical compliance standards. Revocation requests must be honored within 10 business days, and AI disclosure is mandatory in several states—including Texas, California, Florida, Colorado, Illinois, and Utah—often requiring clear identification at or near the start of the call as noted in multiple state-level regulations. These requirements apply regardless of whether the call is managed internally or through a third-party provider, placing the compliance burden squarely on the entity initiating the call. Liability extends to the caller, even when using vendors like My AI Call Center, which operates under strict list discipline and consent verification to support compliant outreach for clinics, franchises, and multi-location businesses.

The Cost of Getting It Wrong

A single non-compliant calling campaign can cost more than most businesses make in a year. That is not an exaggeration — it is the arithmetic of the TCPA, and it applies to AI voice calls just as much as to old-school robocalls.

Under the TCPA, statutory damages run from $500 to $1,500 per call, with no aggregate cap, according to compliance analyses of AI outbound calling. The FCC can pile on top with fines of up to $23,727 per violation, per regulatory guidance on AI voice calls. And the legal environment is heating up: TCPA filings are up 95% year over year, with aggregate class-action verdicts across the docket exceeding $925 million.

The settlements tell the story plainly:

• Gen Digital agreed to pay $9.95 million in January 2026 over prerecorded voice calls to non-customers.
• QuoteWizard settled for $19 million — a benchmark for the upper end of recent TCPA class actions.
• Hy Cite Enterprises paid $4.75 million in early 2026, showing even mid-sized campaigns draw eight-figure legal attention.

Now scale that to a real campaign. A non-compliant campaign of 100,000 calls could expose a business to $50 million to $150 million in damages, based on TCPA exposure modeling. A marketing team that buys a cheap contact list and plugs it into an AI dialer without checking consent records is not running a campaign — it is loading a legal liability onto the company's books.

Here is the detail that catches most businesses off guard: liability extends to the entity on whose behalf the calls are made, even when a third-party vendor places them, as TCPA legal documentation makes clear. You cannot outsource the risk. The burden of proving consent sits with the caller, and it requires records of who consented, when, where, and how — retained for at least five years.

That is why the questions you ask before choosing a calling provider matter more than the price per minute. A provider worth working with will review your list source and consent records before launch, decline lists that cannot support the campaign, and honor opt-outs immediately. My AI Call Center, for example, checks list source and permission records before any campaign runs and flags bought lists without clear consent documentation.

Compliance is not a feature you add later. It is the difference between a campaign that grows your business and one that ends up in a settlement press release.

How to Run Automated Calls the Right Way

Running automated calls effectively means treating compliance as a design constraint, not a checkbox. The FCC's February 2024 ruling confirmed that AI-generated voices are "artificial or prerecorded" under the TCPA, which means every call using this technology triggers the same consent and disclosure rules as a traditional robocall — regardless of how human it sounds. Non-compliance isn't theoretical: statutory damages run $500 to $1,500 per call with no aggregate cap, and FCC fines can reach $23,727 per violation.

A human-in-the-loop model keeps quality and compliance aligned. AI handles the repetitive work — dialing, lead prioritization, voicemail drops, data entry — while humans step in for rapport, objections, and complex conversations. Research shows that teams using this approach see up to 4x productivity gains, and leads are 7x more likely to qualify when contacted within an hour. The key is a clean handoff: when a prospect asks for a person or the conversation exceeds the script's scope, the transfer happens instantly.

• Start with one clear goal per campaign — confirm, qualify, remind, survey, retain, or connect
• Review list source and consent records before launch; decline lists without clear permission
• Approve scripts, disclosures, and escalation paths before anything runs
• Route hot leads to humans for complex conversations
• Log opt-outs and DNC requests immediately and honor them within 10 business days

My AI Call Center structures every campaign around this sequence: goal first, then list and consent review, then script and escalation approval — nothing launches until the client signs off. Outcomes route back into the CRM with disposition codes, per-call notes, and follow-up requests, so the team sees exactly what happened. Calling starts at 9¢ per connected minute with a rate locked for the campaign, and the first campaign review is free.