Misinformation and disinformation are circulating through social media and are a scourge in local and national elections, but now there’s a new twist: artificial intelligence and deepfake videos and audio.
In response, state Rep. Trisha LaChica introduced House Bill 1766, which would require the disclosure of campaign materials that are “fraudulent and deceptive deepfakes of a candidate or party” before a vote.
“The growing threat of deepfakes is just something we need to address immediately,” LaChica says. “It’s a big election year. It’s a presidential year, and it’s really about protecting the integrity of our elections, especially in Hawaii.”
According to HB 1776, California, Michigan, Minnesota, Texas, Washington, and Wisconsin have enacted legislation to prevent the spread of false information through political deep faxes.
La Chica’s bill doesn’t explicitly ban the use of artificial intelligence or deepfakes, but it does have “strict disclosure requirements,” he said. “I think it’s really up to us, the younger generation, to educate the aunties and uncles about these kinds of things because they don’t even know what deepfakes are, right?”
HB 1766 defines deepfakes as “any form of media that has been altered and manipulated to misrepresent someone, usually in a way that makes the person say something. Be what has never been said.”
The House Judiciary and Aeronautics Committee recommended that the measure pass with amendments on February 7.
Two other bills introduced this legislative session would limit the distribution of fake information via deep fax to protect the integrity of local elections.
Senate Bill 2687 would prohibit anyone from “distributing, or contracting with, any materially misleading media to another person unless the media contains a disclaimer.” The bill would also create a minor misdemeanor charge for violators.
State Sen. Carl Rhodes introduced both SB 2687 and SB 2396, a companion bill to HB 1766, calling the measures “an important first step.”
The Senate Judiciary Committee deferred SB 2396 on February 9. The same committee advanced SB 2687 with amendments on February 15.
Rhodes noted that the bills “may not be all that useful” five years from now, “but I think we have to try to respond to technologies as they emerge.”
Ryan Ozawa, emerging technology editor for tech news website Decrypt, said local lawmakers are understandably concerned because it’s “a very new and scary technology”.
But the bill, as written, “is a little absurd, in that I believe that the act being done, which is impersonating another person, is already a crime.”
“It’s like passing a bill that says it’s illegal to stab someone in the eye with a fork,” Ozawa said. “And that’s great. But why a thorn? And why their eye? Are we concerned about causing bodily harm to someone? Then maybe it should be a crime, like not hurting people, rather than hurting people. How to hurt?
When asked if he knew of any artificial intelligence or deepfakes being used in Hawaii for disinformation or political messaging, Ozawa said, “I’m sure they exist. But I’m not aware of any and certainly none that have gone viral to the point where it has affected an election or helped the news.
With the 2024 presidential election on the radar, the use of artificial intelligence and deepfakes “could be extremely powerful and extremely dangerous in politics,” said Colin Moore, an associate professor at the University of Hawai’i Manoa.
“It’s a new technology,” Moore said, “with very little regulation.”
“I think these bills are trying to start the process of regulating this new and very powerful tool of communication,” he added.
Sign up for our free morning newsletter and stay informed every day.