The Senate Rules Committee on Thursday advanced legislation that would mandate a disclaimer in political ads using generative AI to depict fictitious events.

The Senate Rules Committee approved Senate Bill 850 (SB 850) on Thursday, a measure aimed at regulating the use of artificial intelligence (AI) in political advertising.

Carried by Sen. Nick DiCeglie, the bill introduces a requirement for political ads created with generative AI to include a disclaimer when the content depicts actions that did not occur, aiming to prevent misinformation and protect the integrity of elections.

Generative AI, under the bill’s purview, refers to AI systems capable of creating synthetic content such as images, video, audio, text, or other digital forms, based on input data. The legislation mandates that any political advertisement containing AI-generated content that falsely depicts a real person performing a non-occurring action must clearly state, “Created in whole or in part with the use of generative artificial intelligence (AI).”

“We’ve all run political ads in the process of getting here in the Senate. Every one had a disclaimer telling voters who paid for approved or sponsored. It’s how voters know who was responsible for the content in the ad,” said DeCeglie. “This bill simply requires an additional disclaimer for those ads if artificial intelligence is used to depict a person doing something that did not happen and the depiction intends to injure a candidate or deceive regarding a ballot issue.”

The initiative comes amid growing concerns over AI’s potential to mislead voters, impersonate candidates, and undermine electoral trust through sophisticated techniques like deepfakes.

“The proliferation of generative AI use and its outpacing of government regulation has created concern among policymakers about its potentially negative effect on the electoral process,” the bill’s legislative analysis notes.

The committee also approved an amendment, stipulating that the disclaimer requirement applies only if the content was created with the intent to injure a candidate or deceive regarding a ballot issue. Further, failure to include the mandated disclaimer would result in a first-degree misdemeanor.

Committee members raised questions pertaining to the enforcement of the bill’s disclaimer requirement, the interpretation of “intent to injure” in the realm of political competition, and how the disclaimer would be implemented across various media formats.

DiCeglie provided clarifications, stating that the enforcement mechanism would follow the framework already established for political advertisement disclaimers in the state, but with adjustments to address AI-generated content. He further elucidated that for printed materials, the disclaimer must be in bold font; for television and video ads, it should be readable and occupy a certain percentage of the screen; internet advertisements must display the disclaimer without requiring any action from the viewer; and for audio advertisements, the disclaimer should be audibly clear at either the beginning or the end of the message.

“As we enter a presidential election cycle, the fact that we’re going to identify this AI technology, especially when it’s used to injure a candidate, is very important,” DiCeglie said in defense of the bill.

Representing Common Cause Florida, Amy Keith spoke in opposition to the bill, contending that the required disclaimer for AI-generated advertisements is “vague and insufficient.” Keith also stated that the disclaimer merely indicates the use of AI without clarifying whether the content was manipulated or depicts unreal events, thereby failing to inform voters about the potential misinformation they are exposed to. She also noted the bill’s lack of provision for injunctive relief, which would allow for the quick removal of misleading ads.

“The disclaimer that this bill would require is a weak disclaimer and unclear. It just says that the advertisement was created with the use of AI,” she said. “We urge improvements on this bill before it moves forward.”