Why AI Calls Need Written Consent in the First Place

The FCC didn't leave room for interpretation. In February 2024, the agency issued a Declaratory Ruling confirming that AI-generated voices count as "artificial or prerecorded voice" under the TCPA, which means every AI outbound call requires prior express consent — and for marketing calls, prior express written consent. The ruling makes clear that the legal status of the voice depends on how it is produced, not how human it sounds.

The stakes are immediate and uncapped. Statutory damages run $500–$1,500 per violation with no aggregate ceiling and no requirement to prove actual financial loss. Class-action settlements in 2025–2026 have clustered in the $5M–$20M range, and aggregate TCPA verdicts across the docket now exceed $925M. Vague or assumed consent is the single biggest liability — consent does not transfer between companies, and it cannot be buried in fine print or obtained through pre-checked boxes.

The regulatory floor is still shifting. The FCC's proposed NPRM would explicitly require written consent agreements to disclose that the consumer authorizes AI-generated content. Meanwhile, the Fifth Circuit's Bradford ruling held that oral consent can suffice — but only in Texas, Louisiana, and Mississippi. For the other 47 states, written consent remains the defensible national standard. The ruling is jurisdictional; it does not override federal PEWC requirements nationwide.

• AI voices are legally "artificial voices" under the TCPA — prior express consent is mandatory
• Marketing calls require prior express written consent (PEWC) with affirmative signature
• Consent must name the specific business, identify the phone number, and state consent is not a condition of purchase
• Revocation must be honored immediately through any reasonable means
• Records should be time-stamped, version-tracked, and retained for at least seven years

My AI Call Center reviews list source and consent records before any campaign launches — bought lists without clear permission records are flagged and in most cases declined. We tell you plainly if the list will not support the campaign, before you spend anything. The first campaign review is free; the full number is known before approving launch.

The Five Elements Every Written Consent Must Contain

Written consent is the cornerstone of compliant AI-powered calling, especially as regulatory scrutiny intensifies. To be legally valid under the TCPA and evolving FCC guidance, every written consent agreement must contain five specific elements that work together to protect both businesses and consumers.

First, the consent must be in writing with an affirmative signature — electronic signatures qualify under the E-SIGN Act, ensuring digital agreements hold the same weight as ink-on-paper. This requirement prevents ambiguity and provides auditable proof of agreement, which is critical given that TCPA violations can trigger statutory damages of $500–$1,500 per call with no aggregate cap.

Second, the document must clearly name the specific business making the call and identify the exact phone number authorized for contact. Vague references or bundled seller lists fail to meet the "clear and conspicuous" standard, especially as the FCC’s proposed rules would soon require explicit disclosure of AI-generated content use within these agreements.

Third, the consent must state that agreement is not a condition of any purchase, good, or service. This separates genuine permission from coercive tactics and aligns with longstanding TCPA safeguards against bundled consent practices.

Fourth, and increasingly vital, the agreement must reference the use of AI-generated voice or content. While not yet a federal mandate, the FCC’s draft NPRM would make this disclosure explicit in written consent for AI-powered calls, reflecting growing state-level AI transparency laws in places like Texas, California, and Utah.

Fifth, the consent must include an unambiguous opt-out mechanism and honor revocation immediately through any reasonable means — a practice My AI Call Center builds into every campaign workflow by logging and honoring keyword opt-outs like STOP and REVOKE in real time.

Together, these elements create a defensible foundation for AI-powered outreach, transforming consent from a checkbox into a transparent, auditable process that supports compliant, effective campaigns. For businesses navigating this landscape, partnering with a service that validates consent records before launch isn’t just prudent — it’s essential for avoiding costly class-action exposure, which has seen settlements range from $5M to $20M+ in recent years.

Consent Mechanics: Unchecked Boxes, Clear Disclosures, and Easy Revocation

The mechanics of consent matter as much as its content. A valid written consent for AI-powered calls isn’t just about what the language says—it’s about how it’s presented, accepted, and maintained over time. Regulators and courts consistently emphasize that consent must be knowing, voluntary, and revocable, which means the process itself must avoid deception or obscurity.

One of the most critical requirements is the use of an unchecked box that the consumer actively selects. Pre-checked boxes are expressly prohibited because they do not constitute an affirmative action, and courts have ruled they fail to meet the TCPA’s standard for prior express written consent. Instead, the consumer must take a deliberate step—such as clicking an empty box or typing their name—to demonstrate clear, unambiguous agreement. This principle applies whether the consent is captured online, via text, or through another digital channel.

Disclosures must also be clear and conspicuous, not buried in fine print, collapsed menus, or hyperlinks leading to lengthy partner lists. The FCC and multiple law firms have warned that placing key information—like the identity of the seller or the use of AI-generated voice—behind a link or in small type undermines validity. If multiple sellers are involved, each must be explicitly named in the consent document; vague references to “our partners” or “affiliates” are insufficient and increase legal risk. My AI Call Center builds this transparency into its consent workflow, ensuring disclosures are prominent and understandable at a glance.

Equally important is honoring revocation through any reasonable means immediately. Consumers can withdraw consent at any time—by replying “STOP” to a text, saying “I opt out” during a call, or submitting a web form—and businesses must act on that request without delay. Requiring a specific method, such as a mailed letter or phone call during business hours, violates the TCPA’s flexibility standard. Prompt compliance not only avoids penalties but also builds trust, especially in industries like healthcare and franchises where consumer sensitivity to unsolicited contact is high.

Finally, recordkeeping is non-negotiable. Consent records must be time-stamped, version-tracked, and retained for at least seven years—well beyond the TCPA’s four-year statute of limitations. This practice ensures businesses can defend against claims long after a call is made, particularly as litigation trends show rising class-action exposure. Experts recommend treating written, documented consent as the national standard, even in jurisdictions where oral consent may suffice, to withstand scrutiny across circuits and state laws. When consent mechanics are handled correctly, they become a foundation for compliant, effective AI-powered outreach.

Matching Consent Type to Call Purpose and State Rules

One signed consent form does not cover every call you will ever make. Consent type must match call purpose, and mismatching the two is one of the most common ways compliant-looking programs turn into TCPA violations.

Consider a typical scenario: a new lead fills out a form on your website and you call back within minutes. That inbound lead follow-up provides only transactional Prior Express Consent — enough for the follow-up itself, but not for marketing content. If you pivot mid-call into an upsell, you have converted an informational call into a marketing call, which requires prior express written consent (PEWC) you never collected. According to compliance guidance for voice AI outbound calling, informational calls such as appointment reminders and healthcare notifications fall under narrow exceptions that permit no marketing content at all.

The stakes for getting this wrong are severe. TCPA damages run $500–$1,500 per violation with no aggregate cap, and class-action settlements in 2025–2026 reached the $5M–$20M range. Consent also cannot transfer between companies — each entity requires separate authorization, and consent cannot be buried in fine print or obtained through pre-checked boxes.

The Bradford ruling changed the map — but not the math.

In February 2026, the Fifth Circuit held in Bradford v. Sovereign Pest Control that the TCPA does not require prior express written consent for automated telemarketing calls to cellphones — oral consent can suffice. But that ruling applies only in Texas, Louisiana, and Mississippi. Written consent (PEWC) is still required for marketing AI calls to wireless numbers nationwide and for prerecorded calls to residential landlines in 47 states.

That is why written, documented consent remains the defensible national standard. Other circuits and state statutes may still demand written consent, and a single national approach avoids tracking a patchwork of jurisdictions. It also aligns with where the FCC is heading: a proposed FCC rule would require written consent agreements to include a clear and conspicuous disclosure that the consumer specifically authorizes AI-generated content.

State rules layer on top of federal consent.

Even with valid PEWC in hand, several state requirements stack on top:

• Texas, California, Florida, Colorado, Illinois, and Utah impose AI disclosure requirements — Texas requires disclosure within the first 30 seconds of the call.
• Eleven two-party consent states, including California, Florida, Illinois, and Washington, require all-party consent for call recording, so recording consent should be obtained in all cases.
• Florida, Oklahoma, and Washington have telemarketing consent statutes that may be stricter than the federal TCPA.

This is why consent verification happens before launch, not after. At My AI Call Center, every campaign starts with a review of list source, consent records, and calling windows — and if the records will not support the campaign, we say so plainly before you spend anything. Since requirements vary by location, industry, and consent status, clients should obtain appropriate legal guidance before any campaign goes live.

Putting Your Consent Records to Work Before Launch

Before any campaign launches, My AI Call Center conducts a rigorous list and consent review to ensure compliance and campaign viability. This step examines the source of the contact list, verifies the presence and quality of written consent records, and confirms that calling windows align with TCPA quiet hours and state-specific regulations. Lists purchased without clear, auditable permission records are flagged and typically declined, as prior express written consent remains the defensible national standard for AI-powered calls, even amid jurisdictional variations like the Fifth Circuit’s Bradford ruling that permits oral consent only in Texas, Louisiana, and Mississippi.

When consent documentation is missing or inadequate, the campaign does not proceed until the client provides verifiable records or revises the list source. Opt-outs and DNC requests are logged in real time during calling and immediately honored across all campaigns, with updates carried into the client’s master DNC file to prevent future contact. These logs are retained for at least seven years—beyond the four-year TCPA statute of limitations—to withstand regulatory scrutiny, a practice endorsed by compliance experts who emphasize time-stamped, version-tracked consent records as critical for defensibility.

Crucially, no campaign launches until the client formally approves the script, AI disclosure language, opt-out handling procedures, and escalation path. This ensures transparency and alignment with both federal PEWC requirements and emerging state AI-disclosure rules, such as those in Texas requiring disclosure within 30 seconds or Utah’s AI Policy Act. By anchoring every step in documented consent and client approval, My AI Call Center mitigates risk while delivering structured, permission-based outreach that confirms, qualifies, reminds, surveys, retains, and connects—without inventing numbers or compromising compliance.

• Statutory TCPA damages range from $500 to $1,500 per violation with no aggregate cap, making consent verification financially critical.
• Class-action settlements in 2025–2026 consistently fell in the $5M–$20M range, underscoring the cost of noncompliance.
• Experts recommend retaining consent records for seven years to exceed the TCPA’s four-year statute of limitations.