AI scams are becoming more personalized. Here’s how to spot one
AI scams are on the rise, and with recent scams in which a Palm Desert man learned about his own funeral and another in which the caller impersonated Palm Springs police officers for extortion, law enforcement is urging the public to stay skeptical when anyone comes calling for money.
According to the FBI, Americans had nearly 21-billion-dollars stolen in cybercrimes last year.
In the past, scams were much more straightforward. Emails promising riches and a Nigerian prince on his deathbed were the stuff of comedic fodder, and it was relatively easy to laugh off these attempted cons.
These days — with AI personalizing scams — it can be a lot tougher to make out what’s made up.
"AI is being used to scrape the internet and find information about you. Sometimes they'll even get your stolen Social Security number and make it seem very legitimate when they contact you," said Michael Lebowitz, a national security attorney who used to work at the Department of Justice, National Security Division, focusing on digital crimes and cyberwarfare.
Lebowitz is also the author of Second Wave: Inside Al Qaeda's Post-9/11 Attack Plan and America's Secret Effort to Stop It.
AI can impersonate your loved ones — even producing the same voices or fake video of them. Recent scams have preyed on the fears of parents or grandparents.
"Sometimes it'll be like a grandparent or a parent, and then there'll be a college-aged student, and their their voice will be cloned, and the AI will be like, 'Hey, I'm in really big trouble. I need money right now. So what families can do is they can get their loved ones together, and they can come up with a password, like a verbal password. Do not record it anywhere," said Lebowitz. "It's a very analog, retro way of protecting yourself from sophisticated AI. So if all the loved ones have this secret password amongst them, and you get a high, urgently panicked phone call saying, 'We need money right now,' [you can] say, 'What's the password?' And if they hedge and they say, 'There's no time for that,' then it is a scam. Hang up."
Lebowitz has analyzed schemes from all over the world — and he says there are some telltale signs of an AI scam.
One, scammers will pressure you to make a decision right in the moment. Their goal is to get you to hand over money while they have you panicked.
Two, the caller will try to prevent you from hanging up the phone to verify what they're saying. They might even threaten you if you try to do so. The more urgent that phone call is, it's more likely it's a scam. But any legitimate person will let you hang up.
Three — anytime they ask for payment in gift cards or crypto — it's an immediate red flag. Government agencies or utilities will never ask for payment in that form.
Finally — if video is involved, look for subtle inconsistencies in lip, synchronization and micro expressions. You can often see glitches around the eyes, hairline, neck, and jaw. In our News Channel 3 news story, you can see an example of an AI video in which the neck muscle flex in a strange way for a split second before smoothing out.
It's also not just the elderly being scammed in this modern age. Lebowitz said victims are from every age range, and that even very sophisticated people that are falling for cyberscams because AI makes it so much more believable.
According to CrowdStrike, AI has helped hackers increase scams and attacks by 89% compared year-to-year to 2025, because AI has been able to magnify the skills and reach of a range of scammers, from the less skilled to the advanced.
Seeing is not believing — not in this day and age. And AI is just getting started.
"This doesn't even get into more emerging cyber technology like quantum computing, which some people think might be available beginning the 2030s," said Lebowitz. "And if that happens, that's going to be AI on steroids, where any regular computer or network can be hacked and everything's stolen within seconds."
AI is evolving so fast that legislation has not been able to keep up. Lebowitz says legislation on anything digital has always been 10 to 15 years behind — but that it is crucial for lawmakers to take action now and anticipate how AI might endanger our systems and societal well-being.
