Seminole County, Florida – Imagine you are standing in a store and a stranger comes running up to you with a phone in his hand and shows you a video of him stealing a car from a parking lot. Although this video looks completely real, it was created by artificial intelligence.
Melanie Valentine said that’s exactly what happened to her at a West Palm Beach hardware store late last year.
“A young man came running up to me with a phone shoved in my face, yelling, ‘Someone’s stealing your truck! Someone’s stealing your truck!'” Valentine said.
She said the video was compelling.
“He showed me a video of my husband’s truck where I parked it, a guy opens the truck door, gets in the truck and drives away,” Valentine said. “It was very real.”
The man then asked her to follow him outside.
“He was saying, ‘Follow me, follow me, I’ll catch you.’ That’s when I thought, ‘What happened?’ Valentine said. “He was very insistent: Come out. We can get them. Of course, that didn’t make sense.”
[WATCH: Law enforcement warns of ‘growing concern’ over AI prank videos]
That’s when Valentine says she realized something was wrong. Later that month, she learned that the same content creator had targeted several other customers at the same store.
“There have been several incidents at local hardware stores where he approaches people in the aisles,” said Lt. Roy Bevell of the West Palm Beach Police Department. “He approached the man and said, ‘Hey, here’s a video of your wife outside with another person.’
Alexis Martinez-Arizara also showed an AI-generated video of a man being dragged away from a gas station in West Palm Beach, according to a police report obtained by News 6.
This week, the Seminole County Sheriff’s Office announced that Martinez Arizara is accused of pranking one of its deputies at a store in Lake Mary.
“It’s outrageous,” Valentine said after News 6 reported on the arrest. “It’s really courageous to do that to a police officer.”
Valentine says the public needs to know that this type of deception is happening.
“I think we need to let people know that these things are going to happen and are ongoing,” Valentine said. “It’s pretty disturbing and upsetting just because it’s being used for nefarious things.”
How AI makes fakes look real
Greg Gogolin, professor and director of the Cybersecurity and Data Science Center at Ferris State University, used himself as a test subject to demonstrate how AI can be used to manipulate video and audio.
“This was created from one photo and an audio sample,” Gogolin said, showing the News 6 team audio that appeared to be him speaking. “Everything is fake. This is not a video.”
He says that creating a voice clone only requires about 10 seconds of audio. However, Gogolin said it took about an hour to create his sample.
When it comes to video, the barrier to entry is even lower.
“At the very least, you need a snapshot of someone,” Gogolin said. “Now you can even take screenshots. The reason social media is so difficult is because people post so many things. And as long as you have one decent-quality snapshot, you’re good to go.”
Gogolin said AI has been around since the 1950s, but hit a plateau before the breakthroughs made by researchers at Google and Meta and when ChatGPT was launched about two and a half years ago.
“Productivity increases are very fast,” says Gogolin. “It’s actually kind of scary where we’re going to be in a very short period of time, given these new models that are coming out.”
Because of the rapid pace of change, the way Gogolin teaches has also changed.
“If I had taught a machine learning class last year, I would have taught it very differently than what I teach now,” he says. “Now we teach people how to use tools to generate code instead of writing it from scratch.”
[WATCH: Elon Musk deepfake scam costs Florida couple thousands of dollars]
How to spot a fake
Even a year ago, it was relatively easy to spot fakes, Gogolin said. That’s no longer the case.
“Do you look to see if the person blinks? Does he have any quirks?” said Gogolin. “Models have managed to overcome that. So it can be a really difficult thing.”
However, there are still some clues to keep an eye on.
“If you have a video created from a snapshot photo, it will look like that person. But the facial expressions and the way things are generated by the AI will look different than what that person actually did,” Gogolin says. “By the way, if you generate something from a video and create another video, it’s a little different. But you can also see things, like the background, and the audio. This gets harder and harder, but listen to the audio and see if the lips match what’s actually being said.”
When mischief becomes a crime
The Seminole County Sheriff’s Office said the Martinez and Arizara cases highlight “growing concerns” surrounding the misuse of technology. However, the definition of what crosses the legal line remains complex.
West Palm Beach police say First Amendment recording rights limit what can be considered criminal activity when someone is in a public place.
“They’re not asking for any kind of money or provoking a police response, which is what makes it a crime,” Bevel said. “But the fact that they are getting close to people and creating moments of confusion is not good in itself.”
Gogolin says the legal system is struggling to catch up.
“The law always lags behind technology,” he said. “Even when it comes to things like privacy, Europe has much stronger privacy laws than the United States.”
Valentine believes there should be consequences.
“It was really frustrating for me that he continued to do the same thing after what he did to me. There were no repercussions for him,” Valentine said. “But until the law changes, we’re here.”
WKMG ClickOrlando Copyright 2026 – All Rights Reserved.
