Artificial intelligence could play a big role in election races in 2024.
Two Idaho legislators propose a way to protect candidates who are victims of misleading AI-generated media.
On Thursday, Boise Democratic Representative Ilana Rubel and Nampa Republican Representative Bruce Skaug presented their FAIR Elections Act to the House State Affairs Committee. FAIR stands for Freedom from AI Rigged.
“We saw this deployed last week in the New Hampshire primaries where there was that robo-call that went out with Joe Biden saying, ‘I want everybody to not vote on Tuesday.’ I know many of you may not be Joe Biden fans, but it can happen to you,” she said.
“They can send out a video of you on social media or a robo-call, in your voice, with your face, saying, ‘I just wanted to let you know I’ve decided to drop out of the race and endorse my opponent and I hope you will all come out and vote for my opponent next Tuesday.’ There is no limit what they can have you say,” she said.
The bill would give candidates who are victims of so-called “synthetic media” the right to bring lawsuits against the originators, whether they be big media companies or home-based video makers.
“Any mischievous person could pretty much pay $20, get on a website and generate a deep fake of you within minutes,” Rubel said. “The potential for this to undermine our elections, obviously, is staggering.”
The presentation struck a chord with lawmakers who will be running for re-election this year. It spawned a robust debate involving a variety of ‘what if’ scenarios, including this one from Rep. Heather Scott, R-Blanchard.
“I can envision a meme of the health and welfare director saying, ‘Stop getting your shots. COVID is gone.’ Whatever. How is that any different from politicians and why are we giving ourselves something that no one else gets?”
“I think you raise a good point that we are probably going to have to pass 50 bills on AI before the end of next year,” Rubel said. “We are just starting to wrap our head around all the mischief that’s going to come out of this.”
Lawmakers asked Rubel about the penalties embedded in their bill. There’s nothing specific, she said. The courts would determine damages if it’s determined someone is responsible. She warned those might be hard to obtain because elected officials are considered to be public figures; they have a high burden of proof that their reputations have been harmed.
Several legislators said they’d prefer the bill include specific damage figures that could serve as deterrents.
Scott pressed again that she’d like the bill expanded to protect other people, including those not considered public figures. Rubel said she specifically wrote the bill to focus narrowly on candidates running for public office, in part because she felt courts would find the law to be constitutional. She said the court record regarding AI-related defamation cases is slim and says it’s difficult to get convictions in cases involving the First Amendment.
Committee members debated whether to hold the bill until it could be amended. They said there’s still plenty of time in the session to improve and pass it in time to apply to the 2024 election. But others said attempts to draft a perfect bill could kill it for the year. They said they would rather move an imperfect bill to the House. The committee voted 8-5 to send the bill without recommendation to the full House.
