‘Elections Project’ will fight political misinformation with digital identity verification
As campaigns for the U.S. presidential election and down-ballot races begin to accelerate, one university is taking aim at politically motivated deepfakes and the potential misinformation they could spread.
Leaders from Utah Valley University’s Gary R. Herbert Institute for Public Policy and the Center for National Security Studies on Tuesday announced the Elections Project, a pilot program designed to give candidates seeking one of Utah’s four congressional seats and the open senate seat the ability to authenticate their digital identities for free.
“They would be able to verify through a chain of custody that content actually originated from them and wasn’t altered, intercepted or created by a third-party actor,” Brandon Amacher, director of the Emerging Tech Policy Lab at UVU, told StateScoop in a recent interview.
Using a verification tool from Provo start-up SureMark Digital Identity Services, political campaigns for the Utah congressional candidates will be able to verify whether election material depicting their candidate’s image or voice originated from their operation or from a third party —allowing them to track deepfake incidents and potentially stop misinformation from spreading.
Utah, like many other states, this year passed a law that requires political ads that use AI-generated media to include disclosures. If violated, the law allows courts to consider the use of artificial intelligence as an aggravating factor in sentencing.
Students and faculty with the Herbert Institute and CNSS will follow the pilot program and help evaluate the impact AI deepfakes may have on Utah’s elections.
“As government officials, we can do all the work in the world to ensure that our election administration process is sound, but if citizens can’t have confidence in the in the campaign information that they’re receiving, then it’s hard to have confidence in the election,” Utah County Commissioner Amelia Powers Gardner said at a press conference Tuesday.
‘We’re on the precipice’
During the press conference, officials involved in the pilot program showed an AI-generated deepfake video of former Gov. Gary Herbert, for whom the academic institute is named.
“Now if you look at those deepfakes, they kind of look a little bit off. And your sense tells you, especially in the video, your senses are telling you that maybe this is something you can’t trust,” Powers said told the audience. “But that is going to change soon.”
Powers added that audio deepfakes can sway opinions more easily than video deepfakes. Amacher said that’s because audio deepfakes are easier to generate, but harder to spot than video deepfakes. Listeners have fewer contextual clues when listening to a recording versus watching a video, where odd facial expressions, video glitches, or incorrect anatomical or background details could tip off viewers.
“Audio deepfakes are very convincing. But video deepfakes are getting better and better to the point where if you aren’t paying attention, they can fool you,” Amacher said. “We’re on the precipice.”
Amacher said that the pilot project will also extend the verification tool beyond the political domain. In August, Utah voters will also be able to use a free browser plugin to verify candidates’ digital content.
‘Policy always lags behind’
Without more regulations at the state or federal level, or stricter security frameworks, Amacher conceded that the onus to verify the authenticity of content falls on individual news consumers, which, he said, is an “unrealistic expectation.”
“Policy always lags behind technology,” Amacher said. “So by default, a lot of the responsibility does fall on the end consumer, but that’s a very unfair and perhaps unrealistic expectation.”
The Elections Project will run from July until the winning candidates are sworn into office next January, when, Amacher hopes the university compile the data to arm local, state and federal policymakers with the necessary information to pass stricter legislation on deepfakes.
