Broadcasters to take on AI police-like role in FCC’s political ad disclosure plan | Story

In a rulemaking decision that advanced last week along a 3-2 party line, the Federal Communications Commission is considering rules requiring disclosure of any radio or television advertisements that use artificial intelligence to create them. So far, the discussion has omitted the AI ​​police-like role that each broadcaster must take when reaching out to political advertisers.

“In particular, we are not proposing to prohibit or restrict the use of AI-generated content in the production of political ads,” the Federal Communications Commission’s proposal said.

Under the proposal (MB Docket No. 24-211), which is now open for public comment, the FCC is considering rules that would require any station that airs political ads to ask its buyers whether their spots contain AI-generated content.

“If the voices and images we see in political ads are being altered by AI, the public should know,” said FCC Chair Jessica Rosenworcel. “For political ads on TV and radio, the FCC’s new disclosure rules would ensure that this is also the case.”

Specifically, a broadcaster would be required to inform the buyer at the time of the agreement to air a political commercial that the broadcaster is required to make an on-air disclosure for any political commercial that contains such AI-generated content. Broadcasters would also be required to ask whether the commercial contains AI-generated content. If so, disclosure would be required.

“The broadcaster would be required to make the announcement immediately before or during the broadcast of any advertisement by or on behalf of a legally qualified candidate for public office, as well as any advertisement on any topic that contains AI-generated content,” the proposal states. The FCC is also considering whether the announcement could be made immediately after the ad.

For radio advertisements, the FCC proposes an on-air announcement in a “clear, distinct, and intelligible voice.” The proposed disclosure for radio advertisements would read: “The following message contains information generated in whole or in part by artificial intelligence.”

The FCC also proposes requiring broadcasters to include a notice in their online political files disclosing the use of AI-generated content for each political ad, using the same standardized language used in live programming.

“We believe this requirement would help provide greater transparency regarding the use of AI-generated content in political ads, for example by allowing listeners, viewers, and other interested parties to confirm which ads broadcast by a broadcaster contain AI-generated content,” the proposal said.

The FCC says the proposal is about transparency, but the rulemaking — which, ironically, was voted on behind closed doors on July 10 — still has several unanswered questions. They include how the broadcasters must respond if the buyer doesn’t respond to their request and what should happen if the broadcaster is informed by third parties about the alleged use of AI. The FCC has not yet concluded that the broadcasters must reject the purchase, something the broadcasters could not do under current law for federal candidates. Other questions that remain unresolved include whether the disclosures should be required in English or the language of the ad, or both.

The FCC is also considering broadening the scope and applying the requirements to candidate and issue ads embedded in syndicated or network programs, recognizing that the broadcaster would not have direct contact with ad buyers in those situations. One idea is to require broadcasters to ask syndicators and networks whether their programming includes ads that use AI-generated content.

Broadcasters “already overloaded”

Commissioner Brendan Carr calls the idea a “recipe for chaos,” especially so close to an election. He says the plan to potentially penalize broadcasters for failing to disclose violates federal law, which prohibits broadcasters from refusing to sell airtime to federal candidates, even if the ad defames other candidates.

“If the FCC today allows broadcasters to deny access to candidates because they do not disclose the scope and extent of AI-generated content, it is not clear how the Communications Act would allow this,” Carr said.

Commissioner Nathan Simington, who also dissented, said broadcasters were “already burdened with regulatory compliance requirements” and this proposal would only make things worse. He also believes it could cause some buyers to switch to other media.

“I don’t think political spending on broadcasting will decrease as a result of these new requirements. But these requirements provide another small reason to move programming to online and streaming platforms and out of the regulated broadcast space,” Simington said. “This long-term trend is already there, and I don’t think the occasion justifies an additional burden.”