Trump Didn’t Get Arrested (Not Yet, At Least). That Was Just AI.
When former President Donald Trump announced in a scathing Truth Social post that he would be arrested on Tuesday, cities prepared for the worst. He did, after all, call on his supporters to “protest, protest, protest” and “Take our nation back.”
His looming not-so-guaranteed arrest comes on the heels of a Manhattan district attorney investigation into the former president’s role in paying adult film actor Stormy Daniels $130,000 in hush money toward the end of his 2016 campaign for president.
But Tuesday came and went, and the former president was not arrested — sidestepping the much anticipated but rather unlikely perp walk.
Still, somehow dramatic images of law enforcement restraining Trump began circulating online illustrating the potential scenario, and some experts are worried it’s part of an intentional disinformation campaign.
One such post was sent from a satire and parody page on Twitter, which posted on Tuesday: “BREAKING: Donald Trump was just arrested by New York law enforcement.” The tweet included two sham images created by artificial intelligence: One of Trump being restrained by a group of five officers and another in which the former president appears to be falling or resisting arrest.
Twitter later included a note on the post saying “Donald Trump has not been arrested” and citing NBC News and The New York Times. The images stemmed from another user who posted a thread of them on Monday, as Twitter noted.
“Making pictures of Trump getting arrested while waiting for Trump’s arrest,” the thread read.
“As we have all seen, it is so much easier to make fake videos and fake images and make them look really realistic,” Vwani Roychowdhury, a professor at UCLA Samueli School of Engineering, told HuffPost. ”Whether it’s the correct information or socially acceptable information or misinformation, in my opinion, they’re all embedded as part of a narrative.”
How To Spot AI-Generated Images
In a world with rapidly advancing technology, discerning between content generated through artificial intelligence (AI) and real content could be a challenge, but skepticism is key.
The fake images of Trump’s arrest were made using a tool called Midjourney v5, USA Today reported.
“You could visually see if it’s not well-made, you can see the differences, that this cannot be a real scene. But the technology is getting better every day almost, so soon it could be very difficult to tell whether this really is real or not,” Roychowdhury told HuffPost.
Eliot Higgins, the founder of Bellingcat and the poster of the original images, told USA Today that people who come across these images should keep an eye out for key details.
“Generally speaking, you should look out for people with extra limbs or fingers,” Higgins told USA Today. “Text is nearly always nonsensical, and logos and badges are often a mess. If you look at uniforms of the police, you’ll notice they are all slightly different. You might also see clothes blend into each other or have strange textures.”
AI Art’s Potential For Harm
The existence of AI art comes with the possibility for harm.
Brandeis Marshall, CEO of DataedX and a professor of computer science at Spelman College, told HuffPost over email that the image spreading on social media was a prime example of disinformation and an “intentional bad act.”
“Drumming up 45′s supporters in an effort to potentially raise funds or incite another [Jan. 6] insurrection isn’t done by accident. It’s done using a calculated strategy,” Marshall said.
Trump could become the first former president and major presidential candidate to be indicted. And he’s spent the last few weeks spinning the possible arrest to make himself appear as a martyr to his supporters.
“What’ll be interesting to know is how much funds has he raised in that time frame, what has been the expenses accrued by state and federal law enforcement and what are the other news stories not covered because so much attention has been paid to this well-executed disinformation campaign,” Marshall said.
But outside of this instance, Marshall said some additional concerns surrounding potential harm include copyright infringement, lack of attribution and lack of compensation.
Roychowdhury echoed Marshall’s point that such artwork could also be used to advance political causes.
“Any news item, any misinformation, put it in the context of the underlying narrative because that’s really driving it, [but] some people would be wedded to those narratives no matter what you do,” Roychowdhury said.
“It does not really matter whether someone does fact-checking…. Because the damage has been done,” Roychowdhury added.
Midjourney did not immediately respond to HuffPost’s request for comment.