CHARLESTON – With the increased use of images generated by artificial intelligence programs and “deep fakes” used to confuse voters, lawmakers are wanting to place limitations on the use of AI during elections.

The state Senate Judiciary Committee amended and recommended Senate Bill 484 for passage Wednesday afternoon, sending the bill to the full Senate.

SB 484 aims to regulate the use of AI and synthetic media in elections. The bill prohibits the distribution of deceptive synthetic media intended to harm candidates or influence election results within 90 days of an election unless a disclaimer is used.

David Cook, general counsel for the Secretary of State’s Office, told committee members that 19 other states have similar laws governing use of AI during elections.

“Right now, AI, at least as it relates to elections, is unregulated,” Cook said. “This is a disclaimer bill. This simply means that someone within 90 days of an election who uses synthetic media generated by AI to attack a candidate needs to include a disclaimer.”

The bill defines “synthetic media” as “an image, audio recording or video recording of an individual’s appearance, speech or conduct that has been intentionally manipulated with the use of artificial intelligence techniques or similar digital technology.” The synthetic media would have to be shown to manipulate people into thinking the media is genuine when it is not or influence an individual about a candidate.

“If I’m a candidate and I wanted to Photoshop a picture of myself with Donald Trump, that would not violate this law,” Cook explained. “This is a law that protects persons from being attacked by other persons via AI. And finally, it’s made with the intent to injure a candidate or influence the results of an election.

The proposed law states that if synthetic media is used, it must include a “clear and conspicuous disclaimer stating that the synthetic media has been manipulated by technical means and depicts speech or conduct that did not occur.” The bill specifies requirements for disclaimer placement, font size, audio delivery, and inclusion in metadata.

The bill includes exceptions to protect First Amendment rights, libraries, educational institutions, and certain news and media outlets with specific conditions, such as bona fide news coverage or routine publication with disclaimers.

“This (bill) is very narrow. I think some of the bills that were introduced last year were very broad,” Cook said. “That’s why we did the research that we did at the Secretary of State’s Office to ensure that candidates are protected from this type of science, but at the same time not infringing on a person’s First Amendment rights.”

Violators could face a misdemeanor charge and if convicted, receive up to a $1,000 fine along with potential civil penalties. The bill requires proof that the distributor of the synthetic media “knows” that it falsely represents an individual. An amendment offered by state Sen. Ryan Weld, R-Brooke, was approved by the committee to include instances where the person “has reason to know” of the media’s falsity.

Complaints about AI and synthetic media use would be made to the State Election Commission through the Secretary of State’s Office. State Sen. Patricia Rucker, R-Jefferson, asked how it could be determined whether someone violated the proposed law.

“Let’s say I claim that this was used against me, and I claim that this is artificial, I never said such and such. How do you determine that it was artificially created,” Rucker asked.

“There is a way actually when you look at…what metadata is,” Cook said. “That’s the data about the data. And that allows us to track this media as part of our investigation. In other words, it would work in conjunction with our subpoena power going after service providers to find out if they followed the rules.”