Defining “AI-Generated Call”

The draft NPRM proposes to define “AI-generated call” as “a call that uses any technology or tool to artificially generate a voice or text using computational technology or other machine learning, including predictive algorithms, and large language models, to process natural language and produce voice or text content to communicate with a called party over an outbound telephone call.” The element “to communicate with a called party over an outbound telephone call” is intended to tailor the definition to the types of consumer-facing outbound calls covered by the TCPA.

The FCC seeks comment on the proposed definition, including whether it is properly scoped to encompass the types of AI technologies that raise heightened fraud and privacy concerns and whether it addresses those potentially harmful AI uses without hindering or deterring positive AI uses. The draft NPRM also asks what changes to the definition the FCC should consider and how the FCC can ensure any definition it adopts keeps pace with technological advances. 

Proposed Disclosure Requirements

The draft NPRM states the position that called parties should be informed about the use of AI both when consenting to receive calls and also at the beginning of any AI-generated voice message. The proposed rules would add language on AI-generated content to existing TCPA caller disclosures.

Calls using AI-generated artificial or prerecorded voice messages would have to include a clear and conspicuous disclosure that the consumer’s consent to receive artificial and prerecorded calls may include consent to receive AI-generated calls.

Autodialed text messages with AI-generated content would have to include a clear and conspicuous disclosure that the consumer’s consent to receive such messages may include consent to receive AI-generated content.

Calls that include AI-generated voice would have to include, at the beginning of each call, a clear disclosure informing the called party that the call is using AI-generated technology.

For calls that require prior express written consent and include AI-generated content, the written agreement authorizing delivery of such calls would have to include a clear and conspicuous disclosure informing the called party that they specifically authorize the caller to use AI-generated content.

The draft NPRM seeks comment on whether the specific disclosures of AI-generated content add value beyond the current requirements to disclose the caller’s identity and whether there are alternate, less burdensome approaches to achieve the FCC’s objectives. The item also asks whether the FCC should prescribe language for the disclosures or require other visual or auditory signals for these calls. It also asks whether consumers should be able to opt out of AI-generated calls but still receive non-AI robocalls and how the FCC could effectuate this option.

Accessibility Exemption

The proposed rules would exempt “artificial or prerecorded voice calls made by an individual with a speech or hearing disability using any technology, including artificial intelligence technologies, designed to facilitate the ability of such individuals to communicate over the telephone” from the TCPA’s consent and identification requirements. The draft NPRM raises several questions about implementing the proposed exemption, including compliance with specific statutory requirements for calls to residential lines or cellular numbers and with the TRACED Act’s requirements.

The item also seeks comment on whether it could, in the alternative, define “artificial and prerecorded voice” in such a way that it would exclude from the TCPA’s purview technologies that are designed to assist individuals with disabilities communicating by voice over the telephone.

Finally, the draft NPRM asks whether its proposed accessibility exemption can be squared with its prior statements in the Soundboard Declaratory Ruling (“confirming that the presence of a live agent on a call selecting the prerecorded messages to be played ‘does not negate the clear statutory prohibition against initiating a call using a prerecorded or artificial voice’”) and the AI Declaratory Ruling (finding “that AI and other technologies that generate human voices fall within the TCPA”).

Real-Time Call Detection, Call Alerting, and Call Blocking Technologies

The draft NPRM asks how the FCC can encourage the development and deployment of technologies on either the device or network level that can:

1. Detect incoming calls that are potentially fraudulent and/or use AI-generated voice based on real-time analysis of call content;

2. Alert consumers to the potential that such calls are fraudulent or AI-generated; and

3. Potentially block future calls that can be identified as similar based on analytics.

It also asks about the legal, technical, and practical barriers to implementing these technologies, including the costs associated with technologies, how to ensure widespread availability, and how to avoid inadvertent blocking of legitimate AI-generated calls, such as public safety messages.

The draft NPRM raises the concern that real-time call detection, alerting, and blocking technologies could pose significant privacy risks because they appear to rely on analysis and processing of sensitive call data that could be combined with personally identifiable information. Accordingly, the draft NPRM seeks comment on:

1. Existing privacy protections offered with these technologies;

2. Existing federal and state privacy laws that may cover these technologies;

3. Whether gaps exist that the FCC should address; and

4. Whether notice and consent requirements or substantive protections would be more appropriate, should the FCC decide to regulate.

Next Steps

The draft NPRM tees up several questions where stakeholders can help clarify definitions and requirements in the proposed rules, as well as offer insight on the potential scope and nature of regulatory interventions.

The FCC plans to vote on the item during its August 7, 2024, open meeting. If the item is adopted, comments and reply comments will be due 30 and 45 days, respectively, after publication in the Federal Register. 

Hogan Lovells is at the forefront of AI and its impact on nearly every legal domain, including the TCPA. Our experts are helping clients navigate the fast-changing landscape of AI and regulation. For more information on how Hogan Lovells can support your organization, please visit our website or contact our team of experts.

Authored by Mark Brennan, John Castle, Ryan Thompson, Jay Mills, and Ambia Harper.