Auto Loan Fraud Got Smarter With AI – So Has Credit Unions’ Ability to Detect

Auto loan fraud jumped to 24% of loan defaults last year.

Why? AI has advanced to the point it can generate perfect fake bank statements, pay stubs and utility bills that human eyes can’t catch.

Jessica Gonzalez, General Manager of Auto at Informed IQ, joined Sarah Snell Cooke in The Credit Union Connection podcast studio to discuss the evidence she’s seeing: the same utility bill used across 43 loan applications at four lenders. A Valentine’s Day bust-out scheme hitting 10 lenders at once. Fraudulent documents that look completely real because of AI.

“If you’re not utilizing AI to combat AI-generated images, they’re so good that the human eye is not gonna be able to catch it,” Jessica said.

Some credit unions and other lenders are already using AI verification to detect auto loan fraud, reprice loans, or reject suspicious applications. Which means the fraud is flowing downstream to whoever isn’t protected.

“If you’re not using AI technology, you’re gonna get the worst of the worst because other lenders are using AI technology to either catch it and turn away loans or catch it and reprice loans,” Jessica explained.

Current fraud rates on unfunded auto deals average 3-4% across the industry. Some credit unions are seeing 7-8% fraud rates even on funded deals.

What Credit Unions Are Seeing

Early adopters report faster funding, reduced fraud, and better member experiences. The AI isn’t just catching bad actors—it’s preventing good members from getting flagged and rejected.

“You don’t wanna push away good clients,” Gonzalez said. “The technology enables that conversation and that’s really important for credit unions to maintain that personal experience with members.”

Auto loans are credit union bread and butter. But if credit unions aren’t using AI to fight AI-generated fraud, they’re leaving themselves wide open.

Time to catch up.

Verification Without Ripping Out Your Systems

Credit unions don’t need to overhaul their entire tech stack. If you’re interested, Informed IQ already integrates with Origence, MeridianLink, and Defi—platforms many credit unions already use. For credit unions not on those systems, there are no-API solutions that deploy quickly.

The technology verifies documents in real time, flags misrepresentation with 99% confidence, and shows the fraudulent document (redacted) so credit union staff can have transparent conversations with members.

Sarah Snell Cooke: Hey everyone. Welcome to The Credit Union Connection. I’m Sarah Snell Cooke, your host. All the loans have historically been credit unions’ bread and butter. They’re also a fraudster’s favorite playground now, and it’s just gotten worse. 24% fraud increase, according to TransUnion. So why? Because AI tools, obviously, can make it easier to generate fake bank statements, pay stubs, utility bills, all the things you need to confirm a loan and that people will not catch.

Today I am talking with Jessica Gonzalez, General Manager of Auto and VP of Consumer Success at Informed IQ. She’s seen it all: the same utility bill used across 43 different applications, Valentine’s Day bust out schemes, hitting 10 lenders at once, and synthetic identity is so good your mom wouldn’t even know.

Here’s the problem. If you’re not using AI to fight AI generated fraud, you’re gonna be an easy target. Other lenders are catching it, repricing loans, turning away bad deals that are suspect, which means the fraud is flowing downstream to whoever is not protected. So Jessica’s going to explain how credit unions can deploy AI verification without ripping out their existing systems, why 3% to 4% fraud rates are going to be the new normal and how to fund loans faster while reducing risk. Let’s connect with Jessica.

Sarah Snell Cooke: Hello and welcome everyone. I am Sarah Snell Cooke, your host here at The Credit Union Connection. I’m joined live with Jessica Gonzalez today. Welcome.

Jessica Gonzalez: Thank you so much, Sarah, for having me. I’m so excited to be talking with you today.
Sarah Snell Cooke: Yeah, and it’s a very exciting topic. Now, you are the general manager of the Auto and Vice President of Consumer Success at a company called Informed IQ. Can you do a little bit of introduction yourself and the company?

Jessica Gonzalez: Like you said, that is my title. Very long-winded. but ultimately, I am kind of the subject matter expert of auto. I’ve been working in the auto industry for quite some time. And then in addition to that, I run account management, customer success.

Ultimately just wanna make sure that our partners, the lenders, really utilize the technology from beginning to end. It is not easy to automate the loan originations process on top of that, looking at fraud and non documentary verifications. So it’s a pretty complicated, mission so we wanna make sure that our lenders have a really strong partner that understands the industry, how it’s going and, and how it’s being used. A little bit about Informed: We’ve been around from about 2017. We are an AI fintech organization focused on fraud, as well as automating the loan originations process.

Sarah Snell Cooke: So auto fraud is kind of the topic that brought us here together today and how credit unions can respond because credit unions, obviously their bread and butter is often auto loans, traditionally, but there’s also been a 24% increase last year, according to TransUnion in auto loan application fraud. So, why autos? Why now?

Jessica Gonzalez: Yeah. Well, I mean, I think there’s no secret, right, that fraud has always been an issue within auto and within every space. I think that with the technology that we’re seeing, Nano Banana, Gemini, Chat GPT, really being able to replicate images is a big issue.

So certain credit unions either require stipulations, documents for verification of income, and then others just require their credit consumer stipulations regardless, fraudulent documentation, misrepresentation. We see it spiking, pretty much doubling month over month and what that really means is that regardless you’re requiring stipulations or not, maybe your partner is, and you’re not. But ultimately it’s creating synthetic IDs. If you’re not utilizing AI to combat AI and have these automated generated images, they’re so good that the human sight is not gonna be able to catch it.

When I used to work for St. Hyundai Consumer, we had a utility bill that was replicated for eight months. Ultimately, last month we saw a bust out, actually in February, for the Valentine’s Day holiday, there was a bust out and ultimately was 10 retail installment contracts across 10 different lenders. But we also see 43 different applications, across four different lenders, utilizing the same utility bill. So those are the same problems that we saw five, seven, 10 years ago. We’re still seeing them today. And so ultimately it’s just gonna get worse and worse just because you are seeing so many documents that are being replicated and misrepresented online because of these easy ways that you can utilize them.

So we had an onsite and we saw people that were able to generate perfect bank statements, perfect utility bills, and they look and feel real. Ultimately, bank statement providers, payroll providers, they’re actually putting in their misalignments. And so if you think about it, it’s like the old school ways where we try to put watermarks on checks.

We wanted to make sure that checks were not just gonna be able to be fraudulently misrepresented. And ultimately because of that, you’re actually pushing back things that you think are fraudulent, that are not fraudulent. So that’s kind of just compiling the issue, which is technology is getting faster and smarter than our human eyes are not cashing everything. What we do think is fraud is sometimes not really fraud. So you are just kind of in limbo there if you’re not using the technology to catch up with it.

Sarah Snell Cooke: Yeah, for sure. And so, you talked a little bit about my next question, which was how fraud has evolved and obviously AI is helping the fraudsters, but it can also help the credit unions and other lenders as well.

Sarah Snell Cooke: Auto loan delinquencies are up three percent, or have reached three percent, which is high, relatively speaking. And so there’s a confluence of things going on, but also is, I was gonna ask some, is a lot of that fraud, or is a lot of that growth due to fraud? And how is AI helping?

Jessica Gonzalez: So there is, correlation from misrepresented documents to performance. Like I said before, we kind of as an industry we’re like, Hey, if you think there’s misrepresentation, but they’re performing, you’re not gonna wanna go pick up a vehicle that is performing. But ultimately that’s what led to synthetic identity issues. So people were able to create, knowing that it was fraudulent documents, but they were like, hey, it’s still performing, so we’re just gonna let it go.

Ultimately, the way that AI is working is that more and more lenders are utilizing AI technology to understand the direct impact. So when we say that there is misrepresentation, we are 99% confident that the document has misrepresentation on it.

When you have that kind of information, you’re able to adjust your credit risk models. So maybe you would still wanna give out a loan and you’re gonna say, Hey, you’re adjusting what they could have done. Right? So there’s, like you said, it’s a perfect storm. So if you’re not using AI technology, you’re gonna kind of get the worst of the worst because other lenders are using AI technology to either catch it and turn away loans or catch it and reprice loans or readjust their credit model so that they can understand what those are.

Or actually go back and tell the dealer like, Hey, you’re actually giving me 63 utility bills from the same, so what can we do about this? So there’s different kind of actions that you can take so that you can prepare, but if you take no action, that’s really where you’re gonna find the issues. Is that because other people are continuously using the newer technology to catch it?

They’re able to fine tweak, their systems so that they’re actually more proactive versus historically we’ve just really been reactive. We’d say, oh, I think there’s a bust out. I’m not a hundred percent sure I see them having others, but then a month later there would be some credit washing.

So you just really couldn’t know for sure what was happening with an individual or not. But now, with the technology, we’re able to alert people in real time that, hey, on Valentine’s weekend we saw 10 different retail installment contracts with the same identity, the same person, their signature is on there.
So ultimately, if you’re part of that fraud network, there’s without doubt that there is fraud happening. And so I think, if you’re not catching up and like you said in auto, you don’t wanna be first, but you definitely don’t wanna be last. And so you really wanna start leveraging that technology sooner than later.

Sarah Snell Cooke: Yeah, and particularly credit unions who, as I alluded to earlier, they have a significant share of the auto loans in the market. they’re also dealing with, of course, a rise in fraud as we’ve been talking about increasing delinquencies and member expectations. So how can credit unions meet expectations of their members while also fighting the fraud?

Jessica Gonzalez: Yeah, I like to say, the expectation, like you said, like dealers, everyone knows. So then, you go to a credit union expecting kind of that white glove service. That neighborhood experience. You really wanna say that their members are feeling that personal touch. What the benefit of it is that you can have the same kind of fraud protection and then really continue to have that relationship with your members being from a credit union, and that’s really what we wanna do.
We wanna be transparent, we wanna make sure that we’re giving the loans to the right people, giving them the privilege, keeping the rates low for your members. You’re able to do that when you have the technology because ultimately you’re not pushing away bad members because that’s not a good experience.

Like I talked about, if a member is providing to you a true pay stub or a true document, and you’re pushing it back and saying, I think this is fraud, nobody wants to have that experience. So having the technology that really replicates what’s great about our technology is that you can actually see the other document.

It is redacted in real time, but you really have the opportunity to say, Hey, this is actually something that we’re having an issue with. And you can alert them, you can kind of be more transparent with those conversations versus being like a black box, you don’t know what’s going on back there. It’s AI, but you have no idea what’s happening. And that’s kind of think what credit unions really don’t wanna do that. They wanna have that personal experience with the members. So I feel like the technology enables that conversation and that’s really important for the credit unions to maintain.

Sarah Snell Cooke: What are the results that credit unions that are using AI, seeing in their fraud?

Jessica Gonzalez: From a fraud perspective, you can see, like I said, not pushing away good clients. And then the secondary one is a reduction in their fraud. Right now, if we do a retro analysis, we’re still seeing unfunded deals three to four percent and that’s average. So if you, there’s certain credit unions that are seeing about seven or eight percent even on funded deals. So you wanna still make sure that you’re having a reduction to at least industry standard of three or four percent of misrepresented docs.

And then on top of it, if you really think about it, you have the opportunity to expand within the market.
Whenever you feel confident that you’re able to expand your portfolio and not add on risk. That’s what everybody wants. You wanna be able to capture better loans without additional risk. And so you really have the opportunity to not only do that, but also do it faster and more efficient. So you’re in the world of top tier organizations, banks that are much larger. And they have a lot more manpower. So if you’re able to continue funding at the same speed, not increase your fraud risk, a win for win.

Ultimately it goes back into just automating the entire loan originations process, kind of being able to speed up that, where you’re elevating the process. So, dealers are expecting the same type of funding, no delays and clean deals getting expected same day, having the opportunity to really do that within hours now.

Sarah Snell Cooke: Yeah, yeah, for sure. More loans on the books, more good loans, less rod and members are happy ’cause they’re getting their loans funded faster as well as the dealers.

Sarah Snell Cooke: How can credits deploy this kind of AI verification without disrupting their existing systems?

Jessica Gonzalez: We’ve been in the market for a while. We support, all types of lenders. and from a credit union perspective, we have really great partnerships. So we have partnerships with Origence. And, you can utilize informed through their DPA platform. We have partnerships with MeridianLink. We have a partnership with Defi.

So depending on kind of your LOS, kind of your aggregator that, we are already marketed into credit unions and so it’s really available and really focused on that opportunity, like I said, to be able to not only automate your process, but then also be able to elevate kind of your fraud risk.

From my perspective, a really great initiative that there’s already credit unions utilizing us. We have direct partnerships with larger organizations like the Suncoast Credit Union, and they’re deploying us, there through our dealer portal that really elevates kind of the experience for their dealerships as well, eliminating a lot of emails, back-to-back delays and then just kind of cross communication there. So ultimately, we have a ability to implement the way that your technology is kind of already utilized and so we’re really flexible. I think it’s really important to know when you’re buying a technology that it’s not just a one fits all, it’s really gonna be able to fit the needs that you are.

I always tell credit unions, too, what is the problem you’re trying to address? And then let’s make sure that we address that. And that’s really ultimately what my team is responsible for, is making sure that we address that first problem that you’re doing. Is that faster funding? Is that fraud risk? Is that able to integrate really seamlessly?

And then we’re able to get up and running. So we even have no API solutions so that if you want to be able to get up and running quickly, not utilize an API, not a lot of tech and you’re not with one of the LOS providers that we already have an integration with, or maybe you’re in a transition to an LOS, we have kind of a flexibility for all of those different levels of support and technology.

Sarah Snell Cooke: So I always allow my guests the final thoughts. What do you have to say? What do you have to leave our credit audience with today?

Jessica Gonzalez: I think fraud, especially when I talk to credit unions, it seems, it’s definitely a buzzword. It’s out there. You go to conferences and I’m on the conference bracket, we’re speaking all the time, and I feel like there’s kind of the fraud talk. People hear all these buzzwords more like hallucinations and thoughts, and we just did a fraud survey and they was talking about that people are already worried about hallucinations, there’s kind of this fatigue. Going around about fraud where they know it’s a problem. But there also is AI helping or not helping.

And ultimately, if you cut through the chase, you need AI to combat fraud. It is what it is, like it or not, but I think really who you partner with from a technology perspective is really important so that you get the best value. You can also ask those questions. I spend a lot of time just, “Hey, what are you hearing? What’s industry standards? What are the rates like?” And you really need somebody in your corner that’s gonna be able to help provide some of those like inputs and understand, what is happening. And if you’re just providing AI technology to a brand suite of areas that don’t service credit unions or service the niche area that you’re expanding into, it kind of puts you in a disservice.

So I would just say that one, make sure you’re asking all those questions to anyone that you’re doing a partnership with. And then secondly, get your references, understand what other people are doing. You don’t wanna be first, but you don’t wanna be lost. And then third, make sure that, whenever you’re going into, these questions, don’t be worried about all of the buzzwords out there. Just start with your problem and go from there.

Sarah Snell Cooke: Awesome. Well, thank you so much for your time and expertise today, Jessica. Appreciate it.

Jessica Gonzalez: I appreciate your time too. Thank you so much, Sarah.