Trump as a cat warriorWill artificial intelligence decide the US election?
dpa
23.9.2024 - 08:47
A flood of sophisticated deepfakes thanks to artificial intelligence - this has not yet happened in this election campaign. What voters see is often cartoonish and grotesque. But what does that mean?
dpa
23.09.2024, 08:47
26.09.2024, 09:30
dpa
No time? blue News summarizes for you
Despite fears of deepfakes in the 2024 US election campaign, cartoonish and grotesque AI-generated images have dominated social media so far.
These seemingly humorous memes are used to spread political messages and sometimes racist conspiracy theories.
Democrats make less use of AI image generators. If they do, they mainly make fun of Elon Musk and Donald Trump.
Generative artificial intelligence has become increasingly popular over the past four years - and experts had feared the worst in the 2024 US election campaign: AI-generated deepfakes flooding social media and being so realistic that voters wouldn't know what to believe and what not to believe.
But so far, that hasn't happened. AI is actually playing a considerable role in this election campaign. Fake, AI-generated images are regularly doing the rounds on the internet. But what voters are often seeing instead of the dreaded flood of deepfakes is cartoonish and eye-wateringly absurd, so much so that even the most naive of viewers might not take it seriously.
But even these memes can be problematic. Eye-catching AI-generated photos and videos - some of which are meant to be funny - have become a useful tool for spreading false and sometimes racist messages with a clear political charge. And candidates and their supporters are helping to share them on social media.
One example of this is the way Trump and many of his allies have kept alive completely unfounded conspiracy theories that Haitian migrants are stealing and eating cats and dogs in Springfield, Ohio.
Not only have they repeated these claims multiple times - they've also spread related, AI-generated memes. One of them, on Trump's Truth Social platform, showed a group of small cats holding a sign that read "Don't let them eat us. Vote for Trump!" in the air.
Such AI-generated images are new, viral means of promoting old anti-immigration narratives, says Francesca Tripodi, an expert in online propaganda. The memes used to reinforce claims about Haitians are anything but humorous, she says.
"When you have elected officials using this imagery as a way to keep racism and xenophobia alive, that's a huge problem," said the sociologist from the University of North Carolina at Chapel Hill.
Republicans play down significance
Republicans defend the pictures as light-hearted jokes - and by-products of Trump's individuality. "There's a personality culture around Trump that encourages this kind of over-the-top style of communication," says Caleb Smith, a Republican strategist. "The intent is to entertain, not to deceive."
Trump and his supporters aren't the only ones creating AI memes, but they seem to use AI image generators more often than the Democratic side. Some left-leaning users have posted AI images mocking billionaire and X-owner Elon Musk, a vocal Trump campaign supporter.
Democrats also circulated AI-generated photos of the ex-president in handcuffs in reference to the legal proceedings against Trump. But Kamala Harris has so far not tended to amplify AI-generated content. She prefers a visual language that does not require AI generators.
Currently, the Democrat's campaign only allows the use of generative AI as a productivity tool, for example for data analysis, says Harris campaign spokeswoman Mia Ehrenberg. Trump's campaign spokesman Steven Cheung did not answer specific questions from the AP news agency, but said that the strategy had not changed since May. At the time, Cheung had stated in an email that the campaign team was not using tools supplied by any AI company.
When fun turns serious
The use of fake, entertaining and often downright grotesque images for political purposes is nothing new in itself. But AI-generated images, with their hyper-realism, are more powerful than cobbled-together Photoshop images or political cartoons and can draw new attention to a political message.
While some of the images relating to the pets in Springfield were cartoonish and silly, many felt like they were reinforcing a damaging conspiracy theory - to the detriment of a community that has since received numerous bomb threats, with evacuations of schools and government buildings as a result.
The speed and accessibility of generative AI tools make it easy to create outlandish political content that is frequently clicked on and "liked" by many. Anyone with an internet connection can access AI image generators, they are a cheap and convenient way for campaigners to respond to online trends and get a message across.
Disinformation has been around for a very long time in election campaigns, it's not a new problem, says Teddy Goff, the director of digital in Barack Obama's 2012 campaign team, but AI makes it possible to "do these things faster, maybe more persuasive and more targeted".
Trump's warming to AI-generated images contradicts some of his comments in a Fox Business interview last spring. There, he called artificial intelligence "very dangerous" and "so scary" because there was "no real solution" to the issues created by the advanced technology.