Robocalls: consumers hate them, plaintiffs’ lawyers love them, and near-sentient robots stop them so effectively that spam calls are a thing of the past. Wait, what?
The FCC is expected to vote later this month to kick-off a proceeding that would ask how artificial intelligence technologies may be used to protect consumers from robocalls and texts. Headlines about the potential harms of AI dominate the news, but the FCC’s notice of inquiry—a draft of which was released last week—is the first step towards potentially regulating the use of AI to stop the scourge of robocalls. We already know that AI can be used to create realistic-seeming images and voices and are approaching human-like interactions, attributes that could make robocalls far worse. But could these tools also help carriers and regulators identify and block fraudulent calls before they reach consumers? That’s the what the FCC is seeking to find out.
The FCC’s notice first takes aim at defining “AI” for the purposes of the Telephone Consumer Protection Act (“TCPA”). The FCC assumes that its TCPA authority already captures AI to at least some extent, relying on the statutory prohibition on “artificial” voice messages. Beyond that, the FCC points to recent AI definitions from federal statutes, the National Institute of Standards and Technology, the European Union, and states as examples, seeking public comment on the merit of adopting one or more of these definitions for TCPA purposes.
The FCC then asks how AI technologies may be useful for detecting and preventing illegal robocalls and robotexts. For more than four years the FCC has worked with industry to automate robocall blocking on a national scale, including through the development and implementation of the STIR/SHAKEN protocol. The FCC’s orders have also given carriers some latitude to develop additional network-level blocking tools, including by adopting a safe harbor for blocking based on “reasonable analytics.” The FCC’s latest notice asks whether AI-based tools might enhance these blocking technologies and, if so, whether the blocking decisions can be independently substantiated to prevent blocking of wanted calls.
AI tools could also be used by outbound callers to improve compliance with telemarketing laws like the TCPA. The FCC asks whether AI could be useful in making more accurate outbound calling lists, tracking consent from consumers, processing do-not-call requests, and being more responsive to consumers. If such tools are possible, the FCC wants to know how it can structure its regulatory approach so as to not deter these possible beneficial uses of AI.
Of course, the FCC also asks for help identifying the potential for AI to make the problem of robocalls and robotexts worse. Specifically, the FCC wants to know how AI can be used to circumvent blocking tools, create even higher volumes of spam calls, and increase instances of fraud through voice cloning and other techniques used to confuse and manipulate consumers. Such AI-driven calls could, the FCC warns, “defraud consumers, introduce harmful bias, disrupt elections, perpetuate the commission of crimes, or induce widespread panic such as by making false emergency robocalls mimicking the voices of public officials or other trusted sources.”
Finally, the FCC asks what else it can do to address AI technologies. While the TCPA gives the FCC clear authority to make “technical and procedural standards for systems that are used to transmit any artificial or prerecorded voice message via telephone,” how far does that authority go towards allowing the FCC to regulate AI? The FCC wants to know.
Assuming the FCC adopts this notice of inquiry on November 15, a public comment round will open later this fall and provide an opportunity for companies on all sides of these issues—callers, carriers, AI developers, and more—to participate. The team at ZwillGen can help you think through these issues and whether, and how, to participate in the proceeding.
Update (11/15/2023): Today, the FCC voted unanimously to adopt the Notice of Inquiry. We’ll post a copy of the final NOI and comment deadlines when they become available.