Committee debates bill requiring disclosures for AI-generated campaign material

Sen. Janet Ellis presented Senate Bill 25, a disclosure-focused measure that would require a visible notice when AI-generated or "deepfake" material is used in election advertising within 60 days of an election and would create limited remedies for affected candidates, including injunctive relief and a complaint process with the Commissioner of Political Practices.

"This is 1 of those bills where I attract more opponents and proponents, but I'm willing to work with them," Sen. Ellis said, describing the measure as a disclosure bill rather than a prohibition and citing U.S. Supreme Court precedents on speech and campaign finance as limiting a ban on content.

The bill's initial language prompted detailed questioning from broadcasters and internet-service providers. Rhonda Wiggers of the Broadcasters Association said the industry supported certain clarifications but asked to remove the words "or should have known" from language that would impose liability on entities that "knew or should have known" content was a deepfake, arguing there is currently no reliable tool to determine automatically whether content is AI-generated. Wiggers also urged striking a requirement that broadcasters make a "good faith effort" to determine whether paid content is a deepfake, saying FCC rules generally prevent broadcasters from altering political advertising.

Opponents and industry representatives also sought clarification that service providers and intermediary platforms not be held liable when a paid sponsor supplies material and a broadcaster or platform simply distributes it. Senators and staff circulated an amended version (SB 25.2) that reflected some broadcaster-requested changes: replacing verbs such as "disseminate" with the narrower "pay for or sponsor" to target the source of paid content, shortening an audio disclaimer from 10 to 8 seconds, and adding exemptions or limited liability language for broadcasters and interactive service providers. Opponents asked the committee to remove language requiring a "good faith effort" by broadcasters to identify deepfakes and to add explicit language allowing actions against any violator in courts of equitable jurisdiction.

Several committee members questioned how a court or regulator would prove content was AI-generated. Sen. Ellis said the process would likely be "complaint driven" and that the Commissioner of Political Practices has access to software that can assist in identifying synthetic media. The sponsor noted that injunctive relief would require a high burden of proof; she referenced a "clear and convincing" standard in the bill for a judge to grant injunctive relief.

Commissioner of Political Practices Chris Gallus appeared as an informational witness and described how complaints are investigated under existing statutes and administrative rules, and industry witnesses asked for explicit statutory protections for service providers that accept and carry paid material. The sponsor signaled openness to amendments and said she would continue working with broadcasters and other stakeholders before further committee action.