The footage of actual ICE raids is dramatic and disturbing enough as it is. Now, Facebook users are being fooled by AI-generated videos depicting even more extreme deportations occurring in public spaces from the streets to Walmart.
As 404 Media reported on Wednesday, numerous deepfake clips have been circulating on the Meta-owned platform, which changed its content moderation rules after the 2024 election and eased restrictions on political topics such as immigration and gender.
LIFE: Funniest posts from women this week (November 8-14)
One viral clip claims to show Latino Walmart employees being lined up and loaded into a van with “Immigration and Immigration Services” written on the back. One employee was ordered to come forward by an ICE officer wearing a “POICE” badge, but then walked back in line in a bizarre side-step.
Other videos posted by the same Facebook account, “USA Journey 897,” show a McDonald’s worker similarly targeted in an immigration raid and a food delivery man on a bicycle ambushed by ICE agents.
None of these videos are real. They’re full of obvious signs of AI – in the Walmart video, the man’s awkward gait and scattered misspellings show it – but people are still drawn to them.
The clip, in which one man commented, “It’s about time they go to Walmart,” has been viewed 4.1 million times on Facebook, with 16,700 likes, 1,900 comments and 2,200 shares.
Another man, who realized it was AI, said, “I almost believed it until the last guy in the video (lol).” (Indeed, this clip was probably generated by OpenAI’s Sora, which is more sophisticated than any AI we’ve seen before.)
One deepfake posted on Facebook shows a Walmart employee being deported. Another video shows a detained woman screaming for her baby. USA Journey 897/Facebook
Another glossier, more dramatic AI-generated video shows a young woman being deported screaming for the baby she left behind.
Judging by the comments section, more people fell for this. There were some comments like, “How can society tolerate this?” But most people praised what they saw. “LOVE ICE! Get them all!” one woman wrote with a heart emoji.
Life: This is the most infuriating criticism of women – it happens again.
As 404’s Jason Kabler writes, it’s disturbing that an account is spreading such disturbing and dehumanizing videos of immigration enforcement, and that Facebook’s algorithm is rewarding them with views and likes.
Equally disturbing is how many people are taking the bait. Sherry Pagoto, a psychologist and director of the UConn Center for mHealth and Social Media, said one reason for this is that deepfakes are becoming more viable.
Pagoto said videos that evoke strong emotions generate a lot of engagement, as deepfakes often do. In a system where clicks lead directly to revenue, the people who create these videos know exactly how to play it. AI-generated clips that capitalize on polarizing politics or harmful social metaphors are almost guaranteed to bring in big numbers.
Just last week, a racist AI-generated video circulated on social media showing Black SNAP recipients being denied EBT cards and claiming that their SNAP benefits were being sold for profit.
“Once an emotion takes over, whether it’s fear or anger, our ability to scrutinize the truth of the content is diminished,” she says. “This is why misinformation is often designed to elicit strong emotions.”
Sherry Pagoto, a psychologist and director of the UConn Center for mHealth and Social Media, said deepfakes are becoming more realistic in appearance. Illustration: Huffington Post; Photo: Getty
A quick critical look tends to reveal telltale signs of AI, like unsteady hands, misspellings, inconsistencies in audio, and a general sense of the “uncanny valley,” but not everyone is paying close attention.
“If you’re sitting in front of the TV or waiting in line for something and you’re paying half your attention to something, if the AI renders it well enough, it might look real,” said Jen Golbeck, a professor at the University of Maryland’s School of Information Studies who studies the use of AI and social media.
Life: Trump’s “ugly” body language spoke volumes after being booed at a football game, experts say
Obviously, this does not bode well for the upcoming elections. Casey Myers, a professor and director of graduate studies at Virginia Tech’s School of Communication, said given the gullibility of Facebook users and advances in AI, bad actors will likely increase their use of AI-generated videos to exploit political and social discourse.
We saw Russia launch a concerted campaign to spread disinformation and divide the United States during elections, and now even political activists are toying with deepfakes. Before last month’s election, Senate Republicans posted a 30-second video on X and YouTube that took Sen. Chuck Schumer’s real words and used AI to fabricate fake video footage.
Schumer’s fake video included a small transparent watermark with the National Republican Senatorial Committee logo and the words “AI-generated” in the bottom right corner. It can be easy to miss it if you’re scrolling through your feed quickly.
“AI is a very inexpensive way to create political content and ads that resonate with your audience,” Myers said. “Even if these videos are obviously fake, they still fuel a larger political narrative.”
“AI is a very inexpensive way to create political content and ads that resonate with your audience,” Myers said. “Even if these videos are obviously fake, they still fuel a larger political narrative.” staticnak1983 from Getty Images
Golbeck said there is absolutely a risk that these videos could sway opinions during a major election, especially for less-informed or less tech-savvy voters, especially if the videos confirm existing biases.
Golbeck pointed out how the Trump campaign, particularly Vice President J.D. Vance, pushed the fabricated claim that immigrants in Ohio were eating people’s pets, just one of the falsehoods being spread during the 2024 presidential election.
Life: Why are Millennials forced to write “Lol” after all?
“The immigrant cat story was confirmed and disproved, but it never went away,” she says. “Vance has lied repeatedly over the last month.”
What that (and these deepfake ICE and SNAP videos) revealed to Goldbeck is that “there is a large group of voters and politicians and elected officials and members of Congress who don’t care about the truth and want to tell a story that gives them the tone they want.”
“AI-generated video makes it easy to create clips that convey that vibe,” she said. “I think we’re on a timeline where the importance of truth in elections has already been lost.”
Goldbeck said that while technology may help generate more disinformation and spread it more widely, the real challenge is that it is now mainstream for government officials to create and commit to stories with demonstrably false claims, even in the face of overwhelming evidence.
“Voters who accept that do so without worrying about the truth, celebrating that their feelings have been vindicated,” she said.
Related…
Read the original article on HuffPost
