By Credit and Collection News Staff
A Massachusetts auto dealership is facing a proposed federal class action alleging that it used an artificial-intelligence-generated voice to make an unsolicited sales call to a consumer’s wrong number, then followed up with marketing text messages despite the number’s registration on the National Do Not Call Registry.
Cambridge Honda, a Massachusetts dealership, is accused of placing AI-voiced robocalls and sending subsequent texts promoting its interest in buying consumers’ vehicles, according to a lawsuit filed in the U.S. District Court for the District of Massachusetts. The suit, reported Sept. 25, alleges violations of the Telephone Consumer Protection Act (TCPA) and seeks to represent other consumers who received similar contacts.law360+1
The complaint underscores a growing compliance concern for creditors, collectors, servicers and other businesses deploying conversational AI: an AI-generated voice may not be treated as merely a new customer-service interface. Under the Federal Communications Commission’s interpretation of the TCPA, AI-generated voices fall within the statute’s restrictions on calls made with an “artificial or prerecorded voice.”fcc
Allegations Center on AI Call
The proposed class action alleges that Cambridge Honda made unsolicited AI-voiced calls and sent follow-up text messages to people whose telephone numbers were listed on the National Do Not Call Registry. The calls allegedly promoted the dealership’s interest in acquiring vehicles from recipients.law360+1
The “wrong number” allegation is important. A business may believe it is contacting a current or prospective customer, but exposure can arise when a telephone number has been reassigned, entered incorrectly, supplied by a third party, or otherwise does not belong to the person whose consent is in the company’s records.
The lawsuit’s allegations have not been proven, and Cambridge Honda will have an opportunity to respond in court. Still, the case illustrates how automated marketing campaigns can turn a single inaccurate phone-number record into a broader putative-class risk when the same technology, vendor, script or list is used across a large calling population.
Why AI Raises TCPA Stakes
The FCC in 2024 declared that voices generated by artificial intelligence qualify as “artificial” voices for TCPA purposes. That means a company using an AI voice in an outbound call must consider the same consent and exemption requirements that apply to artificial or prerecorded voice messages.fcc
For telemarketing calls to wireless numbers using an artificial or prerecorded voice, the central issue is typically whether the caller had the required prior express written consent. The FCC says AI voice cloning and related AI-generated calling technologies are subject to the TCPA’s artificial-voice restrictions unless the call falls within a recognized exemption or the recipient has given the necessary consent.fcc
The case also reportedly involves follow-up text messages. Text campaigns can create a separate compliance analysis, including whether the sender had valid consent, whether disclosures were adequate, whether opt-out requests were honored, and whether the texts were sent to a reassigned or otherwise incorrect number.
For collection agencies and creditors, the practical distinction between a live agent using AI assistance and an AI-generated voice initiating an outbound consumer call is increasingly consequential. AI may help an employee summarize an account, recommend a next-best action or draft an agent script without necessarily becoming the voice of the call. But when the consumer hears a synthetic or AI-produced voice, the campaign may implicate the TCPA’s artificial-voice provisions.
Do-Not-Call Claim Adds Another Layer
The complaint also alleges that recipients were listed on the National Do Not Call Registry. The FTC advises that businesses generally may not make sales calls to numbers on the registry unless an exception applies, such as a consumer’s written permission or an established business relationship, subject to the applicable rules and limitations.consumer.ftc
The National Do Not Call issue is distinct from, and can coexist with, an artificial-voice claim. A company could face scrutiny over:
-
Use of an AI-generated, artificial or prerecorded voice without the required consent.
-
Marketing texts sent without valid consent.
-
Calls to a number on the National Do Not Call Registry.
-
Failure to identify or suppress reassigned, incorrect or wrong-party telephone numbers.
-
Failure to maintain internal do-not-call procedures and promptly process opt-out requests.
TCPA private actions can carry statutory damages of $500 per violation, with potential trebling to $1,500 per violation for willful or knowing violations. The aggregate exposure can become material in a class action where a common campaign reaches thousands of numbers.classlawgroup
Lessons for Collection and Credit Operations
Although the defendant in this case is an auto dealer rather than a collection agency, the allegations are directly relevant to credit and collection operations experimenting with outbound AI.
First, companies should inventory every outbound communication channel—not just legacy dialers. That includes AI voice agents, vendor-provided virtual assistants, prerecorded-message tools, interactive voice-response systems, ringless voicemail tools, SMS platforms and any technology that can generate or deliver consumer-facing voice content.
Second, consent records must be connected to the specific telephone number called or texted. A customer record alone is not enough if the number has changed hands or was entered incorrectly. Controls should include consent-source documentation, timestamped disclosures, number-validation processes, reassigned-number screening and prompt suppression of wrong-party contacts.
Third, businesses should build escalation paths into AI workflows. If a consumer says, “You have the wrong number,” “Stop calling,” or “I never gave permission,” the system should not simply continue its script. It should capture the statement, suppress the number where appropriate, halt future automated outreach and preserve an auditable record of the request.
Fourth, creditors and collectors should scrutinize vendor contracts. An AI vendor’s assurance that its technology is “TCPA compliant” does not eliminate the caller’s risk. Organizations should determine who selects the audience, supplies the telephone data, establishes consent, controls call timing and content, handles opt-outs, and bears responsibility for errors.
Compliance Takeaway
The Cambridge Honda suit is a reminder that AI does not create a separate, less-regulated lane for consumer outreach. Instead, it may increase litigation risk by making mass outreach faster, more scalable and more difficult to monitor when data, consent and opt-out controls are weak.
For collection professionals, the immediate takeaway is straightforward: before deploying an AI voice agent or automated text workflow, confirm that the organization can document valid consent where required, identify the current subscriber or customary user of the number, honor revocations and wrong-party notifications, and audit each vendor’s role in the campaign. The FCC has expressly stated that AI-generated voices are artificial voices under the TCPA, making those controls central to any outbound AI strategy.
The exchange for the debt economy
