AI-assisted fraud rings are contributing to a rise in losses at financial institutions such as credit unions.  

Fraud losses reported by consumers in the United States rose 25% year over year in 2024 to $12.5 billion, with investment scams accounting for $5.7 billion, up 24% YoY, according to Federal Trade Commission data. 

The jump is due in part to fraudsters becoming more sophisticated during the pandemic shutdown and accelerated by AI availability, Eric Steinhoff, executive vice president of client impact at fintech Scienaptic AI, said during a Nov. 19 webinar hosted by the company. Scienaptic provides AI-driven credit-decisioning and fraud detection technology for auto lenders, including credit unions, banks and other financial institutions.

“Of the identify verification failures we’re seeing, one in 20 … trace back to deep fake attacks,” Steinhoff said. “The volume of deep fake attacks is doubling every six months. This is where a large part of that rapid growth in fraud is coming from. It’s people leveraging AI, creating synthetic identities, building up credit slowly and then butchering us as lenders for large dollar takeaways.” 

“The volume of deep fake attacks is doubling every six months. This is where a large part of that rapid growth in fraud is coming from.” — Eric Steinhoff, Scienaptic AI

Fraudsters’ shift to sophisticated synthetic IDs, account takeovers and credit washing leaves fewer “breadcrumbs,” making it “even more challenging to find a needle in a haystack than it was ever before,” Vinay Bhaskar, chief operating officer at Scienaptic, said during the webinar. 

Lenders are consistently navigating the rise in fraud, Daniel Chiappone, chief lending officer at Space Coast Credit Union (SCCU), told Auto Finance News, a sister publication to FinAi News. 

“We’ve worked closely with our dealer partners. Fraud affects them just as much as it affects us,” he said. “We’ve been doing a lot of training, and engagement from dealers on that front is good.” 

SCCU also works with local law enforcement to identify red flags, which the credit union relays to dealers, Chiappone said. Bust-out auto fraud schemes — in which fraudsters take out multiple car loans within a short period of time and then sell the vehicles overseas before the loans hit credit reports — is also a growing problem, he said.  

“It’s buying 10, 15 cars over a weekend with a straw buyer,” Chiappone said, noting it’s an issue affecting the credit union on indirect lending with dealers. “On the direct side, we see some cars that have been, we think, exported illegally, and the collateral doesn’t exist. We put some measures in place with technology and inspections when there are red flags to mitigate that.” 

‘Sophisticated’ fraudsters at work 

What makes it more difficult for lenders is that fraudulent applications often appear to be from borrowers with high credit scores and income levels, Steinhoff said. 

An unnamed Hawaii-based credit union, for example, was hit by a fraud ring, leading to the financier approving applications for months that seemed legitimate, he said, without providing further details. 

The credit union discovered the fraudulent activity because the ring was using similar addresses and employer names, Steinhoff said, noting that “most fraud is caught at onboarding.” 

“The sophisticated fraudsters, the ones who know what they’re doing, are clean,” he said. “They’re creating clean bureau profiles so that they can get approved and get credit line increases before they abscond with your money.” 

Fraudsters “apply for membership, apply for a small loan — say, like $1,000 credit card or $1,000 personal loan — pay for it for a couple of months, and then come back requesting the limit increase or the increased loan size,” Steinhoff said. “The smart fraudsters have good credit, but they’re also willing to wait it out and increase their potential takeaway from the credit union.” 

Many credit unions are also facing fraudulent information about the collateral, Bhaskar said. 

“The valuation of the vehicle is overstated or the condition of the vehicle doesn’t match what the reality is,” he said. “That feeds into your dollar losses.” 

Fighting AI with AI 

Lenders are turning to AI to fight AI-generated fraud using tactics such as identifying behavioral patterns of applicants and tracking information including how many applications are being received from the same IP address during a specific period, Scientaptic’s Bhaskar said. 

Deeper looks at applications that on the surface meet underwriting criteria may also find that the borrower’s income was overstated based on property records or investment accounts, known as “asset income mismatch,” he said. 

“About half of the applications which are anomalous have found out to be an anomaly of this nature,” Bhaskar said.  

Fraud schemes are also gaining traction on social media, SCCU’s Chiappone said.  

“It’s positioned in a way to people that it’s not illegal, or not really fraud.” — Daniel Chiappone, Space Coast Credit Union

“It’s positioned in a way to people that it’s not illegal, or not really fraud, or this is a way that you can make money,” he said. “We [work] with law enforcement and have been successful in getting criminals arrested.”

Editor’s note: A version of this article first appeared on Auto Finance News, a sister publication to FinAi News.

Register here for early-bird pricing for the inaugural FinAi Banking Summit 2026, taking place March 2-3 in Denver. View the full event agenda here.