The AI Election: Experts warn of misinformation leading up to voting day
WOODBURY, Minn. — As Dr. Manjeet Rege sat at his kitchen table, he flashed a slight grin as he watched an AI-generated video of Donald Trump and Elon Musk dancing to 'Stayin' Alive.'
"Shows you the power of AI," Rege said. "And it shows you some amazing dance moves."
The video, shared to X by Trump Thursday, has racked up 43.2 million views as of Friday afternoon. "Is this real?" reads a top comment.
"If the 2016 election was the 'social media' election, the 2024 election can be termed the 'AI' election," Rege said.
Rege directs the Center for Applied Artificial Intelligence and teaches on AI at the University of St. Thomas.
"It's become a lot easier to generate content that looks very, very realistic," he said.
This is in part, Rege says, to the emergence of generative AI – first with the release of ChatGPT in 2022 and strengthened by developing programs each day since.
"Generative AI has really picked up, to a point where you don't need to be advanced in terms of AI knowledge to create content that looks very, very realistic," Rege said.
It doesn't take long to see AI at work in American politics. Rege pointed out while some are using the powers for good, like using AI to help craft policies that reflect concerns shared by constituents, others are using it to spread misinformation.
After 15,000 or so attended a Kamala Harris rally near Detroit, Michigan, a user on X posted a photo of an AI generated crowd, claiming Harris' crowd was not real.
"That's a great example of disinformation," Rege said. "This is where I think we need better regulation."
Rege says the Federal Election Commission and Federal Communications Commission have had disagreements in how AI should be regulated.
So far, Congress has passed no legislation to regulate A.I's powers. In October 2023, President Joe Biden issued an executive order trying to curtail AI being used maliciously – putting much of the responsibility on social media platforms to self-regulate – which Rege says can be a flawed system.
"Social media companies, they want more content to spread," he said. "They want people to log on. They want more people engaged on their platforms. If you leave it to them to put in restrictions and it's more of an honor system, not enough might be
done."
Rege says leading into the election, there is an extreme need for media literacy to avoid misinformation. Yet as AI improves, he says it could be even harder for the trained eye to discern real from fake.
"People may not have time to really analyze if something is AI generated or not. It could be in the last two or three days (before the election) where somebody who's undecided or someone who's even decided on a political candidate, may see something that's AI generated and quickly change and go on and vote for another candidate," Rege said.
