As AI rapidly develops, electronic scams follow suit

Apr. 18—Artificial intelligence is known for enhancing everyday technologies and improving the efficiency of industries, and while it can be used to help, it also can become a tool to harm.

As the rapidly growing technology becomes more accessible, some are using AI's abilities to assist in the creation of highly sophisticated and targeted electronic scams, from voice cloning to blasting ads on popular social media platforms and websites. The Federal Trade Commission saw a rise in complaints in the past year about ads that either used AI or claimed to use it to steal information and lure people into scams, according to a Bloomberg analysis.

St. Joseph resident Ruth Johnson now largely avoids Facebook because of a situation last year left her scammed out of more than $100 after she clicked on an ad link and made an online purchase on a website she thought was legitimate.

"I got caught up in the scheme," Johnson said. "Just to prevent it from ever happening to me again, I don't do it ... because you can't contact them back because they don't answer."

At least a third of the FTC complaints about AI-assisted scam ads in the last year were spotted on popular sites like Twitter, Facebook and YouTube.

In some cases, customers are lured by ads or posts promoting notable bargains or discounts and then directed to scam shopping websites. There, they end up making purchases for fake products or items that never arrive. In other cases, the links infect a computer with malware, which can lead to personal information and data being stolen, something known as "malvertising."

Facebook officials have identified several large malware families targeting people through ads, browsers and other social media platforms. A FOX News analysis showed some of ads received upwards of a 100,000 downloads.

The ordeal was especially troubling for Johnson because sites like those were places she thought would be largely free of issues. As sophisticated as some schemes have become, she's hesitant to even accept friend requests now after the incident because she's unsure if it could be a fake or imposter account.

"I've heard other people say they got scammed doing it too," Johnson said.

St. Joseph police do not have any confirmed reports of AI-related electronic schemes in town. But officers have heard reports in other locations of highly advanced technology — potentially AI — capable of copying a person's voice and using it for things like opening credit accounts and stealing information.

The technology only requires three seconds of audio to start regenerating a voice.

"It will hear your voice over and over again and your name on your voicemail, on your phone ... they will take that and learn," St. Joseph Police Sgt. Patrick Zeamer said.

Zeamer, a member of the department's crime prevention unit, said highly advanced algorithms allow for more frequent and targeted online scams. Algorithms do not guide AI but rather enable it how to learn to operate on its own.

"The automated stuff where the more you talk about something, if you search one thing, how it ties to everything else, because the algorithms that track those and start showing you all that kind of stuff," Zeamer said. "And that's a lot of that's how you get led down that path."

Staying up to date with developments on unique and powerful technology like AI and potential criminal uses will be important for the department to keep one step ahead if cases pop up locally.

Detective Brendan McGinnis, who investigates electronic crimes, attended a large training convention organized by the Missouri Department of Public Safety this month that covered a variety of topics, one of which was artificial intelligence in law enforcement. The discussion was lead by a New York City police detective who was front and center in working a rise in cases in his jurisdiction.

"He fully understood the (AI) concept, and they started seeing cases of AI pop in New York City. And they decided that, 'Hey, we better do something now before it's too late,'" McGinnis said.

The discussion was one of the first AI-focused learning sessions for McGinnis, providing he and others with awareness and updates on the capabilities of artificial intelligence, which he called an eye-opening learning experience.

"All of them (AI) have guardrails. And those guardrails are to prevent the system from going outside of itself or what its intended purpose was. But those guardrails are about like humans," McGinnis said. "You know, humans make mistakes."

McGinnis said the St. Joseph Police Department is equipped with progressive technology that can allow detectives to make key distinctions between actual audio and video and things that have been digitally altered or generated.

"But we're able to use some of the resources that we have in me knowing the capabilities of those resources to help in this arena that we aren't in yet. But maybe at some point," he said. "The whole concept of it's a self autonomous self-learning, predictable machine makes it just something that law enforcement should be aware of."

AI can create an even more secretive and complex barrier between law enforcement and those responsible for manipulating it, which could make it difficult for authorities to locate and prosecute criminals.

"It's something that we're definitely going to have to stay on top of," McGinnis said. And it really kind of generates a whole 'nother trail that we have to be worried about."

On the other side, the technology could prove to be a beneficial tool for law enforcement as time progresses. The remarkable accuracy of AI could aid in a variety of areas like financial crimes, among others.