News
Fake Livonia murder case goes viral on social media
May 30, 2024
GRAND RAPIDS, Mich. (WOOD) — A video circulating on TikTok got a lot of attention online over a murder story that turned out to be fake and most likely AI-generated.
The video made its rounds on social media, detailing an alleged murder outside Livonia High School. The TikTok, which has since been removed, depicted a mugshot-style photo of a man with the headline, “Revenge Killing at Prom: Man Avenges Son’s Death by Murdering Cop’s Daughter.” A small photo of a police officer and a teen girl is also pictured. The next slide in the TikTok shows a written story about a man who murders a teenager at her prom to avenge his son, who was killed by the girl’s father.
Within a day, the video had millions of views and more than 700,000 comments. Although the original video is no longer online, it’s been reposted by other social media accounts and continues to circulate. It turned out to be a fake, presumably AI-generated story.
There is no news coverage of the case and Livionia High School does not exist in Michigan. News 8 also contacted the Livonia Police Department, which said the incident never happened there.
“The images look real, the images and the story are playing to people’s strong emotional feelings that would trigger strong emotions for such a story,” said Denton Bobeldyk, Assistant Professor at Grand Valley State University’s School of Computing and Information Systems.
Even Bobeldyk, an expert in artificial intelligence, said he had a difficult time deciphering if the images and story were real, which raises questions about how the everyday person would be able to tell the difference. Bobeldyk said it’s even more difficult with text.
“Super difficult to tell with stories, with text data, what’s real and what’s not real just because with text we use the word ‘dimensional,’ it’s very small dimensional. It’s a much smaller space, it’s easier to produce text data that looks real. With images, it’s a little harder, although you just see in the last few years AI has made such strong leaps and bounds in how real the images look. It’s almost, I don’t want to say impossible but it’s very difficult to tell,” he explained.
Earlier this month, TikTok announced it would begin labeling content created using AI to try to combat misinformation, but Bobeldyk said it would be extremely difficult if not impossible to filter out what’s AI-generated.
“I don’t think you’d be very accurate with that. The quality of the videos that are being generated are very good and if someone puts a lot of money into generating something, especially if they have a message that they want to try to tell or try to stir certain emotions or feelings in people, and those are going to be well-funded, it’s very difficult if not impossible to tell.”
Bobeldyk added that especially with social media, videos often gain traction because they stir up emotion and people often share them immediately without thinking too much about it.
“I think it’s going to be on us to sort of do a little due diligence and do some research. Perhaps use some of our intuition if something sounds like it’s too playing on your emotions, or something sounds like it’s not quite real or would Biden do this or would Trump this or whatever it might be, to maybe investigate a little bit further before we share those.”
Article written by News 8's Brittany Flowers. Read original article here.