FCC Proposes Rules Requiring Disclosure of AI-Generated Content in Political Ads

This presidential election season has heightened concerns about the use of artificial intelligence (AI)-generated content in political ads to mislead voters. 

With this concern in mind, the Federal Communications Commission (FCC) has proposed rules to mandate disclosure of AI-generated content in political advertisements on media platforms subject to FCC jurisdiction. In its recent Notice of Proposed Rulemaking (NPRM), the FCC acknowledges that AI-generated content in political ads can be a beneficial tool for election candidates, such as by enabling candidates to more easily create advertisements tailored towards specific communities. The FCC also notes that such AI-generated content could deceive the public, such as in the case of deepfakes. Because FCC rulemakings typically take at least a year to finalize, any final rules adopted in this proceeding would likely not be effective until well after the 2024 presidential election.

In her statement accompanying the NPRM, FCC Chairwoman Jessica Rosenworcel highlighted the concerns, noting “roughly three-quarters of Americans say they are concerned about misleading AI-generated content” and asserting that “the public has a right to know” if AI is used to create political ads. Chairwoman Rosenworcel cites the Biden robocall deepfakes during the New Hampshire primary as an example of the “rising tide of disinformation” and emphasizes the importance of delivering on a “key principle of democracy – transparency.” See our prior blog post for more on the emerging legal landscape surrounding election-specific deepfakes.

The Proposed Rules

• Defining AI-generated content. The FCC seeks comment on how it should define “AI-generated content” for the purposes of its rules. The FCC proposes the following definition:

“An image, audio, or video that has been generated using computational technology or other machine-based system that depicts an individual’s appearance, speech, or conduct, or an event, circumstance, or situation, including, in particular, AI-generated voices that sound like human voices, and AI-generated actors that appear to be human actors.”

This definition would be the first official definition related to “artificial intelligence” adopted by the FCC. Notably, the proposed definition is arguably broad enough to cover not only content that is generated with AI technologies, but also content that was created using more traditional editing tools, such as CGI and VFX software.

• Who the proposed rules cover. The FCC’s proposed rules would apply to radio and television broadcast stations, cable operators, Direct Broadcast Satellite (DBS) providers, Satellite Digital Audio Radio Service (SDARS) licensees engaged in origination programming, and permit holders under Section 325(c) of the Communications Act. Due to the FCC’s lack of jurisdiction, the proposed rules do not apply to political advertising on social media platforms or video and audio streaming services, such as AppleTV and podcasts.
• Intersection with FEC authority. Chairwoman Rosenworcel frames the FCC’s proposed rules as being complementary to the Federal Election Commission’s (FEC) election integrity efforts. The FEC does not have the authority to regulate television and radio stations or advertisements that are not by federal election candidates, such as advertisements by state election candidates or independent issue campaigns.
• Inquiry requirement. The FCC proposed a new requirement that stations airing political ads must inquire whether those ads contain AI-generated content. The stations would be required to inform the party requesting airtime that an on-air disclosure is required for any political ads that include AI-generated content. The FCC seeks further comment on whether such an inquiry would effectively identify all political ads using AI-generated content and if the party making the request would generally know whether AI was used in the ad’s creation.
• Disclosure requirement. The FCC also proposed a new rule requiring covered parties that air political ads containing AI-generated content to make an on-air announcement disclosing that those ads contain AI-generated content. The FCC proposes that the announcement could be made either directly preceding or during the ads using standardized language. For example, the FCC proposes that political advertisements containing AI-generated content that are broadcast on the radio be preceded by: “The following message contains information generated in whole or in part by artificial intelligence.” The FCC seeks further comment on the sufficiency of the proposed disclosure, language requirements, and the appropriate actions for stations to take when informed by a credible third party that an ad contains AI-generated content despite the requester indicating otherwise. Additionally, The FCC seeks comment on how the disclosure requirements should apply to political ads that are embedded within network or syndicated programs aired by broadcast stations, as the station would likely not have direct contact with the party that purchased airtime from the network or syndication company.

Next Steps

This NPRM is the culmination of a months-long effort by the FCC to propose transparency requirements related to AI-generated content in political ads, which Chairwoman Rosenworcel announced in May of this year. Public comments are due 30 days after the NPRM is published in the Federal Register, with reply comments due 45 days after publication.

© 2024 Perkins Coie LLP

Follow us on social media @PerkinsCoieLLP, and if you have any questions or comments, please contact us here.