Commissioners Will Likely Approve AI NPRM With Limited Tweaks

“We’re not expecting significant changes,” an industry official said. The FCC has been active on robocalls, a second official noted, and that will likely continue through year's end and beyond: “Hating robocalls is pretty bipartisan.”

FCC Republicans Brendan Carr and Nathan Simington voted for the notice of inquiry approved in November, which led to the NPRM, though Carr warned of possible regulatory overreach (see 2311150042).

Robocalling is an area where the FCC has clear authority over AI, Carr said at a recent event (see 2407170055). We must find “the middle bowl of porridge” and “commonsense guardrails,” Carr said. The Republicans opposed a recent NPRM seeking comment on requiring disclosures for political ads that use AI-generated content (see 2407250046).

The U.S. Chamber of Commerce and Incompas, joined by the Cloud Communications Alliance, advised the FCC to take a cautious approach on AI and robocalls. The FTC made the only other filing in the docket (see 2407300049).

The Chamber said it’s “premature” for the commission to consider “any regulations" on AI-enabled robocalls or robotexts. The FCC should “continue to monitor how AI technologies develop and then further evaluate both the benefits and risks they pose to consumers, and whether new rules would better protect consumers and enable innovation,” the Chamber said.

The Chamber applauded a February declaratory ruling prohibiting voice-cloning technology in robocall scams (see 2402080052). However, questions remain, including how the FCC will define AI, “which in other contexts has presented challenges given the lack of stakeholder consensus,” the group said. It called for coordination among federal agencies.

Incompas and the Cloud Communications Alliance met with aides to all five commissioners to urge them to reclassify part of the draft, on real-time call detection, call alerting and call blocking technologies as an NOI rather than an NPRM. New technologies “would require significant, costly upgrades to network infrastructure to accomplish the agency’s goal of processing voice communications through AI technologies in real time,” the groups said.

The Incompas-led filing also warned of the potential costs. Enabling the real-time monitoring of voice traffic using AI “would require investments in licensing, bandwidth (up to ten times the current budgeted amounts), data center rental space, and network infrastructure (up to three times the current budgeted amounts),” they said. The FCC should seek more information on potential costs before proceeding to rules, they added.

The cloud alliance is “hopeful” the commission will add the questions the groups proposed in their filing, emailed Brownstein Hyatt’s Michael Pryor, who represents the group. “We are particularly focused on questions around IP interconnection and the impact that the continuing presence of [time division multiplexing] networks may have on the efficacy of any technological solutions the FCC might propose,” he said: The continued presence of TDM equipment “has impaired the effectiveness” of the secure telephone identity revisited (Stir) and signature-based handling of asserted information using tokens (Shaken) protocol, Pryor said. It also “undermined the substantial investment providers made to implement that solution and we would hope to avoid a similar situation with any AI-related measures.”

“AI holds great promise in many aspects of daily life,” the draft argues: “But it also poses considerable potential harms, including facilitating fraud and other deceptions.” The NPRM proposes a definition of an AI-generated call and would require that callers using AI-generated artificial or prerecorded voice messages “include clear and conspicuous disclosure that the consumer’s consent to receive artificial and prerecorded calls may include consent to receive AI-generated calls.”