INDIANAPOLIS (WISH) — Researchers at Purdue University tell I-Team 8 that they are already seeing an increase in the amount of deepfake content on the web as the 2024 presidential election heats up.

“Disinformation, misinformation are as old as politics itself. Here we have the ability to proliferate content that is increasingly credible in ways we couldn’t before,” said Purdue professor Daniel Schiff.

Schiff’s research focuses are artificial intelligence policies and ethics. He is tracking deepfakes online. He showed I-Team 8 some of the tools people are using to create fake content:

• Auto voice generators that will say anything people type into them.
• Picture generators.
• A generator that creates headshots of people that are fake.

“So I can take one of these images or I can take a thousand of them, and I can create social media profiles that are for or against a political candidate,” Schiff said as he pressed a button multiple times, creating a new headshot with every click.

Schiff’s database of content sits at more than 400. Some of the content is completely fake. Some is real but is claimed to be fake, and some is altered to be different than the original. The content is targeting the political spectrum.

“The most concerning thing from a political context would still be the possible use of deepfakes to swing election results; to misinform people in the public about political positions, about electoral processes, to encourage animosity between social or political groups,” Schiff said.

Who are the bad actors in this situation?

“The stereotypical assumption here would be hostile foreign actors, and, indeed, there is a lot of action by foreign parties trying influence elections around the world but also a large portion of misinformation, disinformation, comes domestically. So, it just comes from political parties, from campaigns, from proxies of those campaigns, or independent groups that have a stake in the fight,” the assistant professor of political science said.

Schiff’s wife and research partner, Kaylyn Jackson Schiff, is also an assistant professor of political science professor at Purdue. She tells I-Team 8 that the Indiana legislature has worked on a bill to restrict AI-generated deepfakes around the timeframe of elections, but it didn’t get over the finish line in the last legislative session, so it won’t be in effect come November.

“Many states have short legislative (sessions) that maybe span only a couple of months, so what happens when technology advances in the subsequent months? Well, you have to wait till the next legislative session,” Jackson Schiff said.

In the meantime, it’s up to everyone to critically analyze everything online to determine if it’s fake content.

“Whenever you’re receiving information, especially when if it’s shocking or inflammatory, or perhaps if it confirms your own suspicions — it’s about a politician you don’t like and it’s criticizing them — be careful. That information could be inaccurate,” Schiff said.

To avoid being hoodwinked by deepfake content, the researchers suggest people confirm something is true by going to at least two other sources of news that are reputable.