Katherine Repasky | assistant editor

President Donald Trump posted a video on his Truth Social account on Aug. 30 depicting Kharg Island, a major Iranian oil export hub, being destroyed in a series of explosions.

However, there was no actual military attack. The video was AI- generated.

Less than a week later, on Sept. 6, Trump once again took to his Truth Social account where he posted dozens of AI-generated images. These included depictions of him alongside futuristic military robots, as well as meeting with former President George Washington.

Abigail Hemmen, professor of political science who studies social media and political communication, noted this growing presence of AI-generated content can create uncertainty for online viewers.

“There is inherently this sort of suspicion or this paranoia that this might be fake or this might be an agenda or some kind of attempt to sabotage someone or a community,” Hemmen said.

Strong emotions are powerful

Some voters, like Paige Fairchild, see this manufactured content as a way to both mobilize supporters and raise tensions between political parties.

“If getting you angry will get you out to vote, then that’s what they will try to evoke.” Fairchild said. “Any emotion is interchangeable in this formula, but normally, the more negative emotions drive people to the polls.”

Fairchild, who is a senior political science student at Duquesne, said that it is often difficult to avoid this content.

“I think it’s hard to escape generated campaigns from politicians on social media, whether it appears on your for you page or in the reposts of someone’s Instagram stories,” Fairchild said.

Kaylie Painter, senior political science major, said she has also seen an increase in politically charged, artificially generated images online.

“It’s getting more difficult for sure to tell the difference between what is AI and what is not.” Painter said. “I never know for sure if it’s AI or people just spreading false information.”

While the technology used to generate such images is new, experts note that many of the political strategies used alongside them are not.

“What has always made propaganda or messaging more generally powerful is having really striking visuals, and so I think that is something that AI makes a lot easier as well,” Hemen said. “These drivers of how we feel, they’re still touching on the same instincts we have as people.”

AI systems do not have the capability to determine whether the information they produce is accurate, said Lei Guo, computer science professor at Duquesne.

“People still need to realize that AI is not real intelligence,” Guo said. “The content often seems to make sense because the models are designed for the user to feel happy and satisfied, this is why often people prefer to use it rather than talking to another person.”

‘We need very good regulations’

As AI-generated content becomes more common in politics, researchers are also examining how it will affect the way people receive and filter information.

Sonakshi Bidyadhar, a Duquesne psychology doctorate student and educator, said repeated exposure makes certain information more prominent in one’s brain.

“The more you repeatedly view a certain kind of content, the more it gets consolidated in your mind, in your memory,” Bidyadhar said.

Because of this, political messaging does not have to convince someone immediately to have an effect.

“With social media, the issue is that you have access to your politicians 24/7,” Bidyadhar said. “This constant access then also means that there is a constant inflow of this information.”

Computer science professionals have said the influx of generated media can be attributed to how easily it can be produced.

As indicated by survey responses to a Pew Research study, 47% of Americans say they have little to no trust in the U.S. to regulate the use of AI.

“Moving forward we need very good regulations and teaching methods on AI otherwise it will remain dangerous,” Guo said.

Experts say political messaging is storytelling

Trump’s use of AI-generated content is an example of politicians beginning to use these methods of communication with voters, Painter said.

“With politics, any publicity is good publicity,” she said. “Some politicians are trying to rile people up, but others are trying to get people mad so they want to make a change, which could persuade them to vote.”

Oftentimes, viewers are exposed to this content without having to search for them, Painter said.

“I think it is produced for the people sort of casually scrolling, whoever that might be. Because a lot of these things don’t stand up to intense scrutiny,” Hemmen said.

Psychology experts like Bidyadhar have noted that these posts accumulate a lot of attention because of the human wanting to be a member of something bigger than themselves.

“You would follow what these people are saying or believe in their philosophy, even though you might not totally agree with it on the face of it,” she said. “You’re going to try to act or at least pretend that you do just because you want to belong.”

As AI-generated political content becomes more difficult to distinguish from authentic material, Hemmen said there could be two different reactions from Americans.

“I think a lot of people will either lose a lot of confidence and become jaded or believe nothing is true,” Hemmen said.

Katherine Repasky can be reached at repaskyk@duq.edu