Remember when scams were relatively easy to detect? You’d get an email from a random address, full of misspelling and poor grammar and possibly the mention of a Nigerian prince. Then in 2020, the sophistication of scams evolved faster than we ever imagined as a result of two global events: the rise of generative AI and the COVID pandemic.

The rise of generative AI
Gen AI made it easy for scammers to create realistic phishing emails with perfect spelling and grammar, recognizable logos, and believable disclosures some of which even warned you about scams for credibility! And they didn’t stop there. Scams spread into our text messages, phone calls and social media.

The opportunity with COVID
Around the same time we experienced a global pandemic, we were house-bound, we became lonely, and our only contact to the outside world was digital. Scammers had the perfect opportunity to hone their skills aided by Gen AI. Romance scams went ballistic, work from home scams were born, and scams in general tripled. And it has not slowed down. In 2024, phishing and spoofing was the #1 complaint to the FBI’s Internet Crime Complaint Center ¹.

Photo courtesy of People

Romance with a Celebrity
I’ve heard countless stories while helping banks and credit unions combat these scams. At one credit union, 3 unrelated members were convinced they were in a relationship with, and sending money to, Tom Selleck and his glorious mustache! So I started asking about it when meeting with fraud executives at other banks and credit unions. Gal Gadot, Dan Ackroyd, Brad Pitt, and the list started to grow. One fraud investigator asked a victim how she was convinced it was really Brad Pitt to which she responded, “He sent me a photo of himself in the hospital and that smile was too good to fake.”

Grandma I need help
One day my grandmother called me to ask if my brother is OK. Naturally I responded, “Um, I think so, why do you ask?” She went on to explain that she received a call from him, saying that he was arrested in Mexico and needed bail money to get out of jail. My brother didn’t sound like himself, so she asked why. She says he went on to explain that he’d been in an accident and roughed up by the police. At that point she ended the conversation and called me – smart Grandma! Another time, a friend at work told me she had received a weird call from her sister asking for money. Her sister’s voice was spot on but she was behaving strangely. So she called her niece and learned that her sister had not been the person on that phone call – smart friend! Both of these stories had a good ending, but that’s not always the case. Read more about voice deepfakes and how to detect them with models trained on voice patters – MFCCs with RNNs or CNNs: blog coming soon – subscribe now to be updated!

Resources:
1. https://www.ic3.gov/AnnualReport/Reports/2024_IC3Report.pdf