#11042 min
FRAUDFORWARD
#111Trust is under attack: Inside the world of deepfakes, with Andrew Austin
68 min
What’s up fraud fighters, and welcome back to Fraud Forward!
This episode opens a little differently. And if you are not paying close attention, you will probably believe it was me. It sounds like me, it looks like me, but every bit of it is a deepfake created with artificial intelligence in about twenty minutes for less than five dollars. That is not a future threat. That is today.
Andrew Austin is one of my favorite people to talk to because he does not do hype. He spends a lot of time exploring where emerging technology meets fraud risk. And this is one of the most practical conversations I have had on the show. We are not here to scare anyone. We are here to make sure fraud fighters understand exactly what they are up against and how organizations need to start thinking differently about deepfake fraud and AI voice cloning.
We walk through the full landscape of deepfake technology together, from voice cloning to live face swapping to lip sync. We talk about what each one costs, how convincing each one is, and where criminals are already using deepfake scams today. Andrew breaks down why identity verification is the most vulnerable point for financial institutions right now, why real time detection of AI-generated content is still an unsolved problem, and why fraudsters do not need Hollywood quality deepfakes to be effective. They just need to be convincing enough.
We do not stop at the threat. We spend the second half of the conversation talking about defenses, device intelligence, behavioral biometrics, consortium data, friction design, and why employee awareness and organizational culture still matter as much as any technology investment. Andrew also shares what excites him about where AI is headed on the fraud fighting side, and what he thinks every CISO should already have on their roadmap to combat AI fraud.
Fraud has always been built on manipulating trust. AI does not change that. It just gives criminals a far more convincing disguise.
What you’ll hear in this episode:
- How Andrew created a convincing deepfake in twenty minutes for less than five dollars, and what that means for fraud teams
- The difference between voice cloning, live face swapping, and lip sync technology, and the distinct fraud risk each one creates
- Why deepfake scams do not need to be perfect to work, and how criminals are already exploiting that reality
- Where financial institutions are most vulnerable to AI-generated content right now, and why onboarding and IDV top the list
- Why real time detection of AI-generated video content is still not a solved problem
- How device intelligence, behavioral biometrics, and consortium data work together to close the gaps IDV alone cannot cover
- Why employee awareness and organizational culture are just as important as any technology investment
- What Andrew thinks every CISO should already have on their roadmap to combat AI fraud
- Where AI is creating real operational efficiency gains on the fraud fighting side, and why Andrew is genuinely excited about it
You should listen to this episode if you:
- Lead fraud, risk, compliance, or financial crimes at a bank, credit union, fintech, or neobank
- Are responsible for identity verification, onboarding design, or authentication strategy
- Want a practical, non-sensational breakdown of what deepfake scams and AI voice cloning actually look like today
- Are building or updating your fraud program and want to understand where AI fraud risk fits into your existing controls
- Are new to fraud and want to understand how emerging technology is reshaping what fraud fighters need to know
Episode notes
What you just heard was a scam
The episode opened with a deepfake. Andrew cloned a voice from a thirty second clip, synced it to an existing video, and produced something convincing enough after just one round of adjustments. The point was not the trick. It was the demonstration. Seeing and hearing it lands differently than being told about it.
Voice cloning is cheap, fast, and already being used
A thirty second audio clip is enough to clone a voice. From there, pitch, speed, and variability can all be tuned until the output is convincing. Andrew has heard credible reports of voice clones bypassing IVR-based voice authentication at financial institutions, the kind where your voice is your password. Vendors in this space have evolved, but anyone claiming to be perfect should be questioned.
Live face swapping exists but has real limitations
Live face swapping requires significant GPU power, usually a high-end gaming machine, and works best when the source and target have similar facial geometry. Where it falls down is mismatched face shapes, hair, beards, and build. For that reason it is not as widely used in scams as other deepfake methods. The barrier to entry is higher, and the margin for error is more visible. The more significant risk right now is animating a still image into a convincing video, which is far simpler and cheaper.
Onboarding and IDV are the most vulnerable points
For the past five to ten years, identity verification has relied heavily on document scans and liveness checks. Turn your head left, turn your head right. Andrew makes the case that this approach is increasingly vulnerable because real time detection of AI-generated video content is not yet reliable, and fintechs competing on frictionless onboarding are inadvertently making themselves easier targets. The answer is not just better IDV. It is layering in device intelligence, behavioral biometrics, and consortium data to build a more complete picture of who is actually on the other end.
Fraudsters do not need perfection, they need enough
The creator's purpose drives how much time and money goes into a deepfake. Scamming strangers who have never seen or heard you is easy. Bypassing a voice authentication system requires more precision. The key insight from this conversation is that criminals are not trying to win awards. They are trying to get paid. Good enough is good enough, and the trust vector, the voice, the face, the urgency in a message, does the rest of the work.
Technology alone does not solve this
Andrew's father was scammed out of twenty thousand dollars over the phone, no AI involved, by someone who knew how to manufacture urgency and emotional pressure. The point landed clearly. Fraud teams can invest millions in technology and still lose to a well-trained human scammer if the frontline culture, training, and procedures are not there. Behavioral biometrics and device intelligence are critical. So is a teller who asks why an 82-year-old is withdrawing cash for the first time.
Where AI helps fraud fighters
Andrew is genuinely excited about what AI can do on the defense side. Not replacing investigators, but giving them better tools. Consolidating alerts around related parties, surfacing connections that would take thirty minutes of manual work to find, and bringing everything into a single view so investigators can spend their time investigating instead of clicking between twenty screens. He sees the greatest near-term opportunity in operational efficiency, not automated decisioning.
Key takeaways
- Deepfake scams do not require Hollywood quality. They require enough trust to get someone to act.
- Voice cloning from a thirty second clip is cheap, fast, and already being tested against financial institution controls
- Onboarding and IDV are the most exposed parts of a financial institution's fraud stack right now
- Real time detection of AI-generated video is not yet a solved problem. Layered controls are the answer
- Device intelligence, behavioral biometrics, and consortium data are essential complements to IDV
- Employee awareness and organizational culture are just as important as any technology investment
- AI is creating real efficiency gains on the fraud fighting side, particularly in investigation workflows
- Fraud has always been built on manipulating trust. AI gives criminals a more convincing disguise, not a new playbook
Final takeaway
Today's conversation was not really about artificial intelligence. It was about trust. Fraud has always been built on manipulating trust. AI does not change that. It just gives criminals a far more convincing disguise.
If the opening of this episode made you a little uncomfortable, good. That is exactly how your customers are going to feel if we do not start preparing for this now.
Episode resources & links:
Connect with Andrew Austin | LinkedIn
Fraud & Risk SME
Product Strategy & Solutions, Sardine
Connect with Hailey Windham, CFCS | LinkedIn
Host of the Fraud Forward Podcast
Banking Community Lead at Sardine
Certified Financial Crimes Specialist (CFCS)
2023 Credit Union Rockstar, CU Magazine
Continuous Improvement Award, SAFE Federal Credit Union, 2023
Top 20 Professionals Under 40, The Sumter Item, 2022
Episode transcript
Hailey Windham
00:05
What's up, fraud fighters? Welcome back to another episode of Fraud Forward. I'm Hailey Windham, and today I have, without a doubt, the smartest guest I've ever had on my show, Andrew Austin. He is the most amazing person, and you should be jealous of me because he's my friend and not yours. Let's jump right into it.
Hailey Windham
00:29
Okay, yeah. What's up, fraud fighters? And welcome back to another episode of Fraud Forward. Normally I'd be the one welcoming you to the show, except technically someone already did that. What you just watched, or for our audio listeners, what you just heard wasn't me. It sounded like me. It even looked like me. It even had the eye movements of me. But every bit of it was created with artificial intelligence. And honestly, if you weren't paying close attention, you probably believed it. So that's why we're having today's discussion. Joining me is I I guess the smartest guest I've ever had, but my really good friend, Andrew Austin. Andrew spends a lot of time exploring emerging technology, AI. And how it's changing the fraud landscape. Today we're just gonna be pulling back the curtain on what's possible, you know. Not five years from now, but now. Today. So before anyone gets nervous, this isn't a tutorial. We're not here to teach people how to make deep fakes. But we're here to help fraud fighters understand what they're up against. And how organizations need to start thinking differently. So Andrew, of course, so welcome back to Fraud Forward.
Andrew Austin
01:36
I I must say, Hailey, it's it's great to be here and I really appreciate the opening. That you say wasn't you. But it sounded exactly like something you would say. So thank you for welcoming me back, welcoming me back on the show.
Hailey Windham
01:51
You're so welcome. And of course, I I wouldn't want anyone else to deepfake me but you, Andrew. So I have to admit, even knowing what we were planning, right, seeing myself talking on the screen was a little unsettling. Especially some of those test ones that you sent me originally, where you really I don't even know how I wanna describe it. Where you really pulled out that Southern drawl that was a little exaggerated. And I'm very glad no one else gets to see that. But yeah. So it was a little unsettling, right? But I wanna start there. Walk us through, you know, what everyone just experienced. How much of what they saw was actually me, and how much of it was artificial?
Andrew Austin
02:35
I think it really depends on what you mean by artificial. The video. The source of the video was you. The source of the audio was you. I just use some tools to clone your voice. And then perform a lip sync over the original video. So taking your voice, really just a 30 second clip of your voice is enough to to clone it. And then I can have you say whatever you want to say. If anyone saw the video that I posted on LinkedIn of you last week. You could see that you were talking about competitive pickle brining, and someone with a PhD in goose dispute resolution. It's it's kind of crazy. But you know, I can have you say whatever you want to say. I'm the smartest guest. And just lip sync that audio to the original video. So is it artificial? Yes. But it's also you. So some of the source is you. But I think the scary part about this is that really anyone can take those things that are publicly available, since you speak a lot. We can take your audio. You're on video a lot. We can take your video. And merge those together to make you do or say whatever that person wants.
Hailey Windham
03:56
Yeah, we did a little test also with my youngest daughter. Where you manipulated the video to say that she was my favorite and that she could have a yes day. And she looks at me and she goes, well, you said it.
Andrew Austin
04:09
It's true. It's true.
Hailey Windham
04:10
Like I had to actually do the thing that you manipulated my voice to say. And I yeah, I'm still trying to talk her out of that yes day. But she thinks she's getting it.
Andrew Austin
04:21
Well, I hope she does. I think she deserves it regardless.
Hailey Windham
04:24
Of course, of course. You know, another thing that I found fascinating is that, you know, we we intentionally started the episode this way. I selfishly wanted that wow moment of, you know, freaking everybody out. Because, you know, right, content. But do you think it was important to show people first instead of telling them about it?
Andrew Austin
04:46
I think it's really important to understand what today's technology can do. You know. It's easy to say AI's scary. Or look how easy it is to do this or that, with AI. But seeing, seeing and or hearing that, really kind of hits home. Right. We we we see in the media, we see influencers on LinkedIn talking about, look how easy it is to do a deep fake. And this and that. And I've kind of, I've made some of those posts myself. From an educational standpoint, it is easy. Some of it's good, some of it's bad. But it's you know, it's I think it's really important to see what the technology today is capable of. To know what financial institutions are up against when it comes to identity. And what individuals are up against when it comes to scams.
Hailey Windham
05:39
So so true. And you know, for for someone like me who I I will, you know, say it for forever. I'm not one of these technologically savvy people. Even though my my brother calls me a wizard because he thinks that I do a lot with computers and the internet. He calls me a computer wizard and I'm like, No, if only you knew. But you know, someone like me would think this would take a a lot of time, right? It it surely would have had to take him weeks. I know it would have taken me weeks to have done something like this. But can you put it today's technology into perspective for like, truly, how simple it was. Like you were describing?
Andrew Austin
06:19
Yeah, I mean it's it's it's super simple. And like when you you look at the evolution of technology over time. I think people are always gonna push the boundaries of what can be done. What should be done. Anyone's that anyone that has ever watched an an action movie over the past several decades has seen visual effects, audio effects. I remember I grew up in the like The Fast and the Furious kind of era. And was big into the car scene. So I loved all the movies. And I remember one of the movies was filming when Paul Walker passed. And they ended it with like a a deep fake essentially of him. Like they had his brother play him. But they superimposed his face over him. So it looked like him. And it was like, wow, that's that's really cool, right? But with today's technology, like it's not that difficult. Like we, we saw and heard in the opening seconds of the show. I can download a video of you. Clone your voice. Create a new video of you that looks and sounds like you in about twenty minutes and costs less than five dollars.
Hailey Windham
07:29
So without giving away the the recipe though, how difficult is something like this in twenty twenty six compared to even, you know, two to three years ago? When you were giving your Fast and Furious, I was thinking Marvel, you know. And I'm like, well, super Superwoman back in the day, you know. The the old, older videos. You can see that it. There's like an actual blink, and then and then, something appears, right? That, but now it, you see it instantaneously. Like it it's not like a, now magic happened. Or even like the genie back in the day. I Dream of Jeannie when she would, you know, and blink her eyes or squint her nose.
Andrew Austin
08:05
You are not old enough to have watched that show.
Hailey Windham
08:09
Listen, I am I am cultured. Whenever it comes to to movies and music. Okay. I know it all. Any genre, okay? But you know, when I think of like how far we've come, right? But we're not talking about that. We're talking just a couple of years ago to to now. So so how difficult is something like this?
Andrew Austin
08:28
Yeah, well, I mean, if if you think back a couple of years ago, we had Will Smith eating spaghetti. That's kind of been the the gold standard for for AI like video generation. And it looked terrible a couple of years ago, right? And we've seen it progress over time to to the point where it looks like him actually sitting on a beach eating spaghetti. I I what I'll say is deep fakes are easy and cheap to make. It doesn't mean that it's simple to make a convincing deepfake. It, something that you need to consider is like what's the purpose of it? So are you trying to convince a family member that you're in trouble and you need money? Are you trying to scam randers, random strangers on the internet? Like the creator's purpose kind of drives how time intensive or resource intensive, money intensive it is to create those deepfakes. You know, I can animate an image of someone you know, in a couple seconds. I can create a video of you in a couple minutes. It really depends on like how convincing I want it to be and what the audience is. So if I'm trying to scam people on the internet that don't know what I look or sound like, that is very easy. Right? We've we've we've talked about pig butchering on this show before. You've talked about it a lot. You've been involved in Operation Shamrock. Like it is, if if you want to move into like a, an operation like that from the scammers side, I can very easily create a digital avatar. And have that avatar do or say whatever I want. You know, there are creators on on Instagram and Twitter or X, Only Fans, that are completely AI generated. And millions of people follow them. So are are are you just trying to make content and gain a following? Like, what are you trying to do with it? Right. In in the same vein, like if I'm a scammer, I can create one of those avatars or one of those people and do whatever I want with it. I can make it do whatever I want, say whatever I want, wear whatever I want. And it is that easy to convince people. Like if you don't know, if you've never seen me before, if you've never heard me before, I can create all the content I want. That looks realistic and make you think that I'm real. I think that's the scariest part of it. You know. The AI is not intrinsically evil. Technology is not intrinsically bad. It's it's what we do with it. You know. Since since computers, since you're a computer person. You know.
Hailey Windham
11:13
Wizard.
Andrew Austin
11:14
I, my family says that I work in computers too. You know, since computers came out. And technology has come out. People always try to push the boundaries. Whether it's hacking a system or. You know, my my father was a was an Oracle DBA. And and did some stuff with SQL or SQL that Oracle didn't even know could be done. Right. So people are always pushing boundaries. What whatever technology it is, they're pushing boundaries to see what can be done with this. And I think, you know, we we see in the news. Like I think who was it? Ben Affleck sold his AI video company to Netflix for I think billions of dollars. Millions or billions of dollars. Right. It can be used for good. It can be used for for movies. This this lip sync that I I created for you for the beginning of the show could be used. You know, if you're creating training material. And you have an hour-long training video that you produce. And you now need to train someone in, let's say, Korean or Japanese or whatever language. You don't need to hire someone that speaks that like language natively. To recreate that video. You can now translate the video or translate the audio and dub over it. So instead of just having subtitles or someone speaking over your lips moving as they were originally. You can dub over that. So it looks like you are speaking Korean. So it it has valid uses. I think the scary part of it is just when people misuse it. And attempt to get around controls for financial institutions, for banks, for scamming others.
Hailey Windham
13:01
The only thing playing in my head right now is.
Andrew Austin
13:07
I I didn't catch that.
Hailey Windham
13:09
So the old Japanese movies. Where they go, or you know, that were converted over into English. And that, you you hear them say Godzilla but then you see their mouth move. So it's like that that's kind of what's going through my head. It's like. We're we're just, technology has to innovate. We get that. I I think that, you know, the the issue that I have and I love that you first started talking about the creator's purpose. I think that that is just like obviously, or we would hope, that these, the people who are creating these technologies aren't doing this with malicious intent. They're doing it because, hey, this is a cool new thing that can be done. And so, but the problem, right, and and that's coming up in fraud, is that technology doesn't stay exclusive very long. You know, you mentioned that that the tools are are accessible. But I'm wondering like, why are they just becoming so widely available? Like even for cheaper. And then, you know, because it's not a Hollywood box office just using it. So it's not just for the the media to be able to, you know, make us really cool movies like Fast and Furious or The Avengers or a New Godzilla movie where everybody talks and you can see their lips moving the right way. I mean, is this it simply better AI, cheaper computing, open source development? Is it all the above? Like, why is it so widely available?
Andrew Austin
14:31
I mean, I think it's kind of all of the above. You know, people want to use new technology. As I mentioned earlier. Like people want to kind of push the boundaries and see what's possible. And that's that's kind of how I got started in this. And you know, it's what can I find? What can I do with this? And in some of the videos that I made for you last week, the lip syncing was kind of bad, honestly. It was, you know, your your mouth wasn't moving right, your your teeth were enormous.
Hailey Windham
15:00
Enormous. Large horse teeth.
Andrew Austin
15:04
It just it just looked very, very strange. Right. So when you put anything out there and it doesn't produce the desired results, there are always people that are going to think, I can make that better. Or how can I, you know, change my prompt? Or, or find a new tool that's going to produce a better output? And I think, you know, if you look statistically, probably the the vast majority of the use cases for for AI are legitimate. You know, you see it in every business now. Two years ago, people were terrified of it. And then everyone rushed. We need AI. We need AI everywhere. Right. So it's just people pushing to to innovate. To automate. To make things better, more effective, more efficient. Regardless of the industry, regardless of what you're trying to do. If you know, if I do something manually that takes me three hours a week. And I can build something, or prompt something, that cuts that down to 10 minutes or zero. Why not?
Hailey Windham
16:07
Yeah.
Andrew Austin
16:08
So people are always going to try to, to make their lives better in some sort of way. I think that's you know, that's why it's out there. That's that's the purpose of it. And most people are attempting to use it for good. I think because of the industry we work in, we see a lot of the bad.
Hailey Windham
16:24
Yeah. Yeah.
Andrew Austin
16:25
We see a lot of the bad. I mean, that's that's all we see, right? It's all we see. No no one is I mean, we we we talk about using AI and AI agents and this and that to to do different things within fraud operations and fraud detection. But you go to a conference, the vast majority of stuff you're hearing about AI is the AI that's being used for for bad. So that's what we tend to focus on and that's what we hear about. We don't think about the positive use cases always.
Hailey Windham
16:54
Yeah. So true. And and I'm excited to talk about the positive use cases a little bit later on. So I'll say to to wrap up this this first part, obviously the the scary part isn't that, you know, you fooled me, right? But it's just that you could. How many people you could actually fool with this technology into thinking, I mean, the one thing is that I think that celebrities are probably like, good. Now I don't have to actually do those cameos. We can just do my little AI avatar and I can say, you know, happy birthday to whomever. And I then I don't have to do it. But I also think that's where this conversation can get really important. Because what we just watched is only one version of what's possible. So, you know, deep fakes have evolved. When most people hear the word deepfake, they they picture, you know, one thing. But that's really become an umbrella for several different technologies. And I I love that you, whenever we were talking about this episode, you you specifically said, why don't we talk through the different categories of like what these are? So I'd love to walk through those now. I think each one creates a different type of fraud risk. So we'll start first with voice cloning. You know, I think we heard a little bit, but how realistic have AI voices become?
Andrew Austin
18:11
I think you heard a little bit.
Hailey Windham
18:12
Yeah.
Andrew Austin
18:13
You've seen how realistic they've become. You know, I I I literally use a 30 second clip to clone your voice.
Hailey Windham
18:20
Yeah, scary.
Andrew Austin
18:21
And I can't, it it'll give me this basic voice, right? I have you. And I can, I can change the speed, I can change the pitch, the variability, I can change different the aspects about the voice. And kind of fine-tune it to get it where I like it. And the cost of this is incredibly low. Is incredibly low to do this. And it's you you you hear how good it is. And I can just keep regenerating, regenerating. Tuning some dials, tweaking some knobs until I get it where I want it to be. Right? So it's it's it's super simple. And it's it's, it can be very dangerous depending on what it's used for.
Hailey Windham
19:02
Yeah. Yeah. What do you think makes our our brains just trust a voice so quickly? When when we hear someone on the other end of the phone? Or whatever it's like. What do you, what do you think?
Andrew Austin
19:13
Yeah. Well, I'm not a neurologist or a psychologist. But I think,
Hailey Windham
19:17
Excuse me, that's under the role of fraud fighter. So yes, yes you are.
Andrew Austin
19:21
Kind of. I could have made you say that I had a PhD in neurology or or psychology.
Hailey Windham
19:25
Next time.
Andrew Austin
19:27
Maybe I'll do that for the next one. However, I think you know, when I pick up the phone and you're on the other end. Or my kids are on the other end. Or my parents are on the other end. I, I am, my brain knows that voice. And I know to trust it. I know how you talk in different situations. I know emotionally how I respond to things. And sometimes you can tell something's off if it's AI generated. But it's it's got to the point where it's it's so good, you you may trust it. Right? I think that's, that's the dangerous part of this. But but your brain just, you you you know the voice, right? So, if it's, if it's close enough, you're going to trust it. And the the danger in that, is that if I establish that I am you on, a let's say I call your husband as you, right? And I say, Hey honey.
Hailey Windham
20:25
Well, we we already, he he listened to the voice. And he, the only only comment he said was, it sounds like you might be live at like NASCAR or something. So I mean just use that. And say, hey babe, I'm here at NASCAR.
Andrew Austin
20:37
Yeah. Yeah. And then we went back and tweaked it. And he said that's the best one yet. So that's, so that's it. So I could I can test these kind of kinds of things out. I could call my dad, my my kids, my whoever, right? And kind of test those out and see, does this pass the muster? And do they really believe it's me? and make it so they believe it's me. And what what do you know if if someone close to you calls you in an emotional state. What what is your immediate response? Like I call you and say, Hailey, I'm in trouble. If all I say is, you answer and say, hey, what's up? I say, Hailey, I'm in trouble. What's your response?
Hailey Windham
21:12
What do you need?
Andrew Austin
21:13
Yeah. What do you need? How can I help?
Hailey Windham
21:17
Urgency.
Andrew Austin
21:18
You create urgency and you create that need to help someone. It doesn't matter what I say next. You're more inclined to do that. If it's totally out of character for me, like if I if if my next statement is I'm in Botswana. And I've been attacked by, whatever, water buffalo. And I need you, just wire me five thousand dollars so I can fly home and get medical treatment. It's absurd, right? But if I say like, hey, I'm in Omaha. And my corporate card's not working, I can't get a hold of anyone. You know, my kids are in the hospital and I need to get home immediately. My wallet's been stolen. Can you buy a ticket for me? You may be inclined to go ahead and do that for me. Right. So it's it's not, it's the voice. It is the urgency. It is the message. That is the call to action. And and that's that's really all that goes into it.
Hailey Windham
22:14
Yeah. you know, where are you already seeing criminals use the the clone voices today?
Andrew Austin
22:21
I, you know, vendors don't like to talk about where they fail. So it's it's really tough to get substantiated claims of yes, it can bypass this. But I have heard, you know, there are a lot of well, not a lot. There are a handful of vendors that do voice recognition like within an IVR system. So you call your bank. And I know the the famous use case is like my voice is my password. And you say that, and that authenticates you. That authenticates me as Andrew Austin. I don't need to give them my social security number, my address, my mother's maiden name, and all those out-of-wallet questions that irritate the hell out of everyone that gets asked them. Right. You know, I'm calling from the same number at the same time, about the same account, blah, blah, blah. Right. I don't want to answer all those questions. So it's a, it's a, you know, a a good customer experience. It's a frictionless, my voice is my password. Right. I I have heard that there are voice clones that can get around those types of things. I think that's the, in financial services, that's probably the main use case where where criminals or fraudsters are are using. Or attempting to use those sorts of things. I will say I'm sure that the vendors that provide those services have evolved. I would highly question anyone that says they're perfect. But you know, it it is a a way that criminals are testing to get around controls with with AI clones.
Hailey Windham
23:48
So true. So obviously the voice alone in itself is powerful. But now let's add a face. So live face swapping almost feels like science fiction, except it isn't. And I got to witness this live at the ACFE conference. And I was trying to bring up the the post so I could see exactly. So I I watch soups live on a panel. Like live face swap with another guy that was on the panel. And it blew me away. I was like, this is crazy. The one thing that like obviously the the internet in the one conference room was a little lagging, but I got to see it like in real time in the like speaker room away from everybody else. And it was so real that I mean even the the guy who it was, he was like, my gosh, it looks exactly like me. Like they had different facial hair, right? But it still looked like the same guy. So, can you explain what live face swapping actually is?
Andrew Austin
24:48
Sure. So there there are a lot of different technologies that that can do this. The barrier to entry for this, I would say, is the power of your machine. So these things really require a lot of GPU power. And, which isn't really found in most corporate or personal laptops. You typically have like a gaming machine that's really powered up. It might cost a couple thousand dollars to be able to run this in real time. And be seamless, right? So I I have this technology on one of my work laptops. Which is not a gaming rig by any means. And if, if I I have a face, like a picture of a face, that is of similar facial geometry, similar facial hair, it looks decent, right? But it doesn't stream in high quality. So it would lag. It it has nothing to do with internet connection. It's it's all about the the power of the machine. So it's all happening locally on your computer. And it what it does, a a live deep fake, it it overlays the face on top of yours. So I've I've I've done this with with many, many people to show. And I was at a conference in Oklahoma City a couple of weeks ago, and did this for some people there. But where it falls down is if I try to deep fake you. Or, you know, I I'm a large man. If I try to overlay someone's face that is significantly skinnier and has a different face shape than mine, you're kind of gonna see some ghosting and some outlines that that don't work, right? Or I have a beard. They don't have a beard. Now it's if I deep faked you, for example. All it does is kind of overlay from the forehead down to the chin. And across to the ears. Right. So if I deep fake you. Which I'm not going to because it would be utterly absurd. It would be Haley with a beard and short hair. So what the the the scammers, fraudsters, whoever that are are doing this. Are are picking images of people that have a similar facial geometry. Maybe they're potentially wearing wigs. They're they're doing different things to augment the technology beyond just the technology. I can't just take a picture of you, overlay it on me, and it looks perfect with hair, glasses, makeup, and facial hair or lack thereof. Right. So it's something I don't think it's used as widely in scams. We we hear about, you know, there's landmark cases of, you know, multi-million dollars lost due to corporate, I think it was over in Asia. the CEO called. Had a video conference with someone and scammed them into sending a huge amount of money somewhere. Right. But it's it's not that widely used. It's not a caveat. It's not as widely used as some other things. Because the complexity behind it. The cost and the complexity of it is is much greater than a simple like animating a a an image of someone. With a source video. It's it's a lot more difficult to to do and it's a lot more it's even more difficult to do it correctly.
Hailey Windham
28:22
Right. [Ad Break (28:29): Finally, I'm so happy to share with you all that the Fraud Strategist is now a podcast. What? Yeah, on top of my weekly newsletter, you could now listen to and watch me talk about my, and hopefully your favorite topic, fraud strategy. And from time to time, I'll be hosting operators and founders to discuss where the industry is headed and what we providers should pay attention to. I must say, I'm super excited. And if I'm being honest, a bit nervous about all of this. I've been debating with myself whether to start a podcast for ages, but kept putting it off. But now this is a result. So I guess there's no turning back. So if you want to join me for the ride, head over to Sardine's website and subscribe now. Are you ready? Am I ready? We'll find out next Saturday.]
Hailey Windham
28:29
Yeah, so I'm gonna share my screen. Hopefully this works. I had a credit union friend reach out to me. I say friend, I mean he deep faked me. But he took a picture of me that I posted on LinkedIn. And then did the face swap with him. And you can see some of the things that you were mentioning, where it's like it's not my same face shape because the beard kind of messes it up, kind of makes me have an elongated jaw. But so this is what it looked like when he did it.
Andrew Austin
30:00
Yeah.
Hailey Windham
30:01
Yeah. So for those that are just in listen only mode, like you can see he's truly like every every movement he makes from you know his head from side to side to lifting his hands, you can see like it it's me. Like my my picture is doing exactly what he was doing. So it it's a little weird to to see it like that. Obviously it wasn't perfect. And that's even like so, we talked like face swap. But that's even like the image to video, right? Of you know, how does a single picture suddenly become someone speaking on camera? Like it it's weird. I I do like the technology that you used for for the deep fake for the original for for the beginning of this call or episode. But, and I think you've already mentioned, right? Of where like the fraudsters are misusing this capability. Where they are showing up on calls. And then the one thing that you mentioned on the first one that people, I don't think think about because I know I didn't, where it's the lip syncing. The the lip syncing technology is is pretty cool. I mean I, it a, once we finally figured out the teeth part. And and gave me normal average size teeth. At least for my face shape. Like it it looked a million times better. So it if you wanna, you know, talk even through the the lip sync aspect of the of the technology, feel free.
Andrew Austin
31:33
Yeah, yeah. So this is actually something I just found yesterday because as we were talking through what we would do for this podcast, I was watching some of the videos that I created. And they were all animating images of you. So when you animate an image of someone, like if the the AI doesn't know how your mouth moves. Or how you emote. Like how how, what do your eyes do when you speak? Like how how do you express yourself? How does your face express what you're saying? Right. So it it's kind of hit or miss. It can be very. It can be close to good. I have, I have,
Hailey Windham
32:15
You have to tell it. Is it more Jim Carrey expression wise? Or or the clear eyes commercial guy? Like where do we sit with expressions?
Andrew Austin
32:24
Yeah, yeah. I made, I made one that I felt was like pretty decent. The one I I posted on LinkedIn was probably the worst one. Because I didn't want to, like, put the best thing out there first, right?
Hailey Windham
32:37
Right, right, right. It was the teeth.
Andrew Austin
32:40
But yeah. So as we were talking through this, I found. I just was, you know, researching. I was talking to Claude, which is another AI tool. On how to, you know, how can I best sync audio to video. And it it recommended a few options to me. And I checked checked them out and found one that offered a free test. And I was like okay, let's go with this. And it was, it was incredible. Like, I was I was shocked. When I, because I was not expecting it to be very good. I took, you you had a a video on YouTube, from a few years ago, that was very high quality. I I downloaded that. And, you know, created the whatever I wanted the voice to say and overlaid it. And it was and it was, different camera angles as well. So it kind of cut halfway through. And there's like cars. You're you're sitting in front of a window and there are cars moving in the background. And all of this stuff is maintained. So you still see the cars moving in the background. You see the camera cut, and it is it is incredibly realistic. How it synced the audio to your lips. So the, like, I think one of the downfalls of that is, like it, you're, if you're speaking with your hands, right? It it may, the timing may be off. But it's still, it's, it can be incredibly convincing. So if you, if I have a video of of a friend of mine. You know, standing somewhere and saying something, in just a contemporaneous conversation, that I take of them. In theory, I could just recreate their voice. Have them say whatever I want to say. And dub it over with lip sync. And, you know, make a new video. And again, this can be a funny thing. But when you think about the fraud aspect of it, it's scary. There are the legal ramifications of it. Right? You know, you're in a you're in a lawsuit for whatever. Or you're going through a divorce. And your spouse says, they said this. And they produce video evidence of you saying something horrible to them. Or to the kids. Or whatever. That is a very scary, very real possibility that could happen. And that's that's kind of where we get in this pickle of, you know. We have these tools that can be used for all sorts of good or creative purposes. But they can also be used for bad.
Hailey Windham
35:17
Yeah, I think that the one consistency we see with these technologies is that, they they do look different, right? If they can attack your eyes, your ears. Meaning like, we're deceiving, being deceived with our eyes and our ears. But every one of them attacks the exact same thing. Which is trust.
Andrew Austin
35:36
Absolutely.
Hailey Windham
35:37
So I think the one misconception that people have is that criminals are, you know, trying to win Academy Awards with this, right? They're not. They're they're just trying to get paid. And you know, I think that we've pretty much covered it just with the quality and the how much it cost. But that fraudsters don't actually need Hollywood quality deep fakes anymore, right? Just because of that, that trust vector. And, you know, we talked about voices earlier too. It's like, we used to say on the front line, well, have you ever talked to this person? Have you ever met this person? Well, the way that FaceTime is now, I mean, I've got a friend that lives clear across the country. She's in Vegas. I'm in South Carolina. We FaceTime so much that I feel like I don't miss her because I see her all the time. It it it's a real relationship. But, and it's someone I've met physically in person multiple times. I've gone and hung out with her. I know this is a real person, just for anybody who's concerned.
Andrew Austin
36:33
Are you sure?
Hailey Windham
36:34
I am sure. I am sure. But you know, there are others who they meet someone online. And they do have these FaceTime conversations all the time. And so for them, it is real. Right? It doesn't have to be perfect like whomever you're trying to clone. Because you just gotta make it look right and sound right. To sound like just someone completely new, right?
Andrew Austin
36:57
So I'd like to share a video that I created using a virtual avatar. So we're talking about, you know, all we need to do is make it look real. And and maybe this isn't super realistic. But this is a video that I created with a virtual avatar. Where I just say, hey. I can have it do, say, where again, whatever I want. So if I have a scammer, and I just want to make someone believe, and I think the prompt I gave for this was like have this man walking down a country road. At the end of the day. With the sun low in the sky. Talk about account takeover. So let's watch this.
Virtual Avatar
37:43
How are you protecting yourself against account takeover risk?
Andrew Austin
37:46
Now it's honestly an absurd video, right? Because the, it, there's no one gonna be walking down the road. But the point is, that I can create a character. And I can just have them do whatever I want. I wanted that to be more cinematic. Just to kind of test of what you could do. But I can have them in a TV commercial. I can have them sitting on a beach. Couple months ago, I using some of Google's tools, created a virtual avatar of myself. And the outcome of that is like, it's me. You know, what what I do, I hold my phone up, I look left, I look right. I don't even think I look up and maybe I look up and down. And then on the screen it says, read these numbers as they appear. And I read 28, 33, 55, 87. And, that simply, I can create a virtual avatar of myself. And then I just prompt it to say, you know, make a clip of me sitting at a at a table in a Miami cafe, drinking coffee, talking about account takeover risk. Or synthetic identity risk. Or whatever I want, right? And the out, the the the output of that is incredibly high quality. It looks like me. And just from a few, a few numbers that I read, it can clone my voice. And and it creates something just absolutely, I mean, believable. I and I put some prompts in to make me look like I was a bodybuilder. Which was slightly less believable. But you know, the the voice, the the movement, everything just looks real. So it's very easy to do this stuff. It's very easy to make things convincing. I think that's what people need to understand. Is that it's it's not coming, it's here. And what we think through as fraud practitioners is, okay, how is it being used. You know, but five years ago, and even still today. A product person, business person comes to us as risk professionals. And says, hey, we're launching this new payments product. Or we're launching this new onboarding. What is the first thing we think about? How is that going to be exploited? How how are we going to like when Zelle came out. Let's go back a decade, right? I remember being at a bank at the time Zelle came out. And we're, you know. We need a couple million dollars for this real-time fraud monitoring platform. And the thought is, well, a couple million dollars. Like how bad could the fraud be? And that's probably the worst product from a fraud standpoint, launched in a very long time. Billions of dollars lost in real-time payments. So, that is instantly what we think about.
Hailey Windham
40:38
Yeah
Andrew Austin
40:39
So when we think about AI, again, it's here. What can it be used for? In a bad way. Right? So how is it gonna be used against financial institutions? How's it gonna be used against individuals? And what can the fraudster or scammer gain from that?
Hailey Windham
40:56
Yeah. And then there's also like the speed versus perfection mindset as well for the fraudsters. Like, would they rather have something that's perfect, or something that's believable enough and ready in five minutes, that they can, you know, exploit the vulnerable there?
Andrew Austin
41:12
Yeah, the the the one of the sayings we we have in fraud is you know, you don't need to be the best. You just kind of outrun the bear, right? You you don't want to spend ten million dollars on your fraud platform. If you can spend five hundred thousand, and then they attack your competitor because it's easier. Right? So it's not a race, you know. I always talk about this in compliance too. Like it's there's no one racing to be the most compliant. Yeah, I don't want to spend fifty million dollars on my compliance program if the auditors and regulators don't have a problem. So I'm not going to innovate in those spaces unless there is a problem. Same in fraud. Like businesses are not going to spend millions and millions and millions of dollars on on fraud prevention if it doesn't make a, make business sense.
Hailey Windham
42:00
So true. So let's let's bring this back to financial services, financial institutions. Where do you think banks are most vulnerable?
Andrew Austin
42:09
Onboarding.
Hailey Windham
42:10
No. God, yes.
Andrew Austin
42:11
You know, for the past five, ten years or so. For the past five to ten years, I think people relied, or businesses relied heavily on IDV. So scan your license front and back. Take a selfie. Move your head side to side. Move it up and down. Move it around in a circle. Move it this way, that way, close your eyes, open your mouth. Do all these different things. Right. It worked largely for a long time. And I think the vulnerability lies in those. Real time, I'll say why. The real time detection of AI generated content is difficult. The technology to, to look at a streaming video in real time, and detect artifacts with no lag and deliver instant results, is just not there yet. There are plenty of companies that do this in near real time. But when you have fintechs and and financial institutions that are fighting for market share, and I can go to one that offers instant onboarding, or another that takes five minutes to give me a decision. And may ask me to, you need to do this this process again or provide this documentation. Who are you gonna go to? You're gonna go to the one that offers the least friction. And the fraudsters are going to go to the ones with the least friction. So I think the the IDV industry is, is particularly vulnerable to this. And people are going to need to rethink, businesses are going to need to rethink, how they verify identity. And I I don't want to take this down a track of talking about identity and a, you know, a single authorized source of identity. Or tokenized identity. Or any of that. I'm just, I I want to make the point that this is probably the most vulnerable place for financial institutions when it comes to AI generated video content. Because it's so easy, as you've seen, to take an image. And just turn your head one way, turn it the other. I can take an image of you, record a video of myself doing whatever the prompt is for that IDV. And in five seconds I've got a video of you. Turn your head left, turn your head right. And if that's all it takes. Then, I'm not saying that every IDV company will fail all the time. and many are very, very good at this, but as the AI tools develop more, they're more and more vulnerable to that happening.
Hailey Windham
45:09
For sure. So I do think though that the good news is, and what I'm I'm picking up is that organizations, they aren't powerless. Right?
Andrew Austin
45:17
Not at all.
Hailey Windham
45:18
Defending against this requires thinking differently than than what we have in the past. So I'd love it if we could shift and start talking about our defenses. You know, we can spend the rest of our time talking about preparation instead of panic. So if you're leading fraud right now, Andrew, where do you start?
Andrew Austin
45:38
To defend against AI?
Andrew Austin
45:42
Well, if we take this from the onboarding perspective, since that's what we're talking about, like IDV, I don't think is enough, right? Maybe it's a a it becomes more of a passive onboarding. Where you allow them to do some things. While keeping strict controls around how much, how much money you can fund the account with. Or when and how much money you can move at what point in time. While you're doing some background processing of the video to look for those AI artifacts. I, that's one place I would would focus on. And two, I think identity has to be, go beyond just a a document or driver's license, and a picture or a video of the face. Right? Consortium data has to play into this a lot. Device intelligence has to play into this a lot. So I live in Cincinnati, Ohio. If I'm onboarding from California, maybe I'm on a business trip and talk to, someone has a cool new fintech or product and I want to try it out. Right? And so you see my IP or my device in California. Well, is that high risk? I don't know. Right. You have to look at it in totality. So look at the velocity of applications. So has this device, identity, whatever, been used a lot lately? You know, if you if you're seeing 10 different applications at 10 different fintechs over the past two days, for a single identity. Then hey, maybe this is odd. Or maybe this person works in fraud and is just trying to test their defenses. Cause I've been known to maybe open an account or two just to see what the onboarding process is like. Right. But the typically, you know, you you need to move beyond that. Right? It's gonna be a fraudster. It's it's it's someone using multiple devices to access an account. Or the same device, same identity, across multiple institutions. So consortium data comes in to this to the picture here significantly. And and understanding, you know, are they changing aspects of their PII? So is it now at a different address? Is it maybe the SSN is different? Or maybe they've changed their birth date, right? All those, all those synthetic identity triggers, if you will, can be looked at to see if that is a true identity. Is it an identity being stolen, abused, or is it synthetic?
Hailey Windham
48:10
How important are simple controls? Like callback procedures and out-of-band verification, do you think?
Andrew Austin
48:18
I'm not a huge fan of like, out-of-wallet questions. Because, actually I hate them. And I know there are industries built around this. But there are hundreds and hundreds of data brokers. That I can sign up for. That are either free or very low cost. And find most of that information anyway. So, you know, I I have signed up for these, I have I've tested them, been verified, is a very common one that people use. I can pull essentially a background check on you. And find out every place you've ever lived. Every phone number you've ever had. Every neighbor, every, you know, pretty much everything about you. Right? Maybe I need to dig a little deeper and find what car you owned in 1999. When you were in high school. Or in your case, probably in elementary school. Or middle school.
Hailey Windham
49:16
No comment.
Andrew Austin
49:17
But you know, it it is it is very simple to to kind of bypass those types of questions. And and and I apologize. Remind me the first part, so I can respond correctly.
Hailey Windham
49:28
Just call back procedures.
Andrew Austin
49:30
You know, I I think, it it really depends on your your volume of, in like what institution you are. You know, if if you're a credit union that onboards like ten people a day or a week. Yeah, call them back. There's there's plenty of things you can do. It it really depends on the size of the institution, the volume that they have, and what specific fraud pressure they have. To design that that fraud program around.
Hailey Windham
49:56
So true. What role do behavioral analytics play now?
Andrew Austin
50:01
A a huge role, I believe. You know, I used to run, well I still am involved in all the of the POCs that we do here at Sardine. And I I've seen a lot of device intelligence and behavioral biometrics PE tests that we do with our our clients and our prospects. And we, if it's a financial institution that has onboarding, we no doubt will always find mass onboardings from a device farm. And when you look at like the orientation of a device or the typing, swiping, or lack thereof, copy paste activity, or scripts running, or is remote access running? is there a bot? Is there this? Is there that? It always pops up. So if you're a financial institution that does a large quantity of of onboarding volume or transaction volume, you're always going to see, I take that back. Typically going to see, those types of activity. With, within your onboarding. Within your transaction flow. Where you have like a single device accessing multiple accounts. Or you have multiple devices accessing a single account. You're gonna see bots, you're gonna see all kinds of different stuff, that device intelligence and behavioral biometrics is is critical for and should be a key staple for any fine financial institution.
Hailey Windham
51:28
I, so true. And it's one of the reasons why I'm such a big fan of Sardine. And why I agreed to come on anyways. Is because the device and behavioral biometrics, just it it was something that obviously coming from a smaller organization, at first I was like, this tech is too much. Like we can't handle this kind of tech. But now, seeing what it can do. And again, just the way that we can get into that proactive state. And where, you know, that friction doesn't have to happen, because we have these other things in place. It was just a game changer. So I'm, I'm glad you mentioned that. But the one thing that I don't want listeners to walk away believing, is that technology alone solves this issue, right? So I'd love it if you could even talk about why employee awareness and organizational culture still matter just as much.
Andrew Austin
52:22
Yeah, it matters maybe even more than technology. Because humans are always going to be the weak link in in a fraud program. Or any program, really. You know. You can, you can invest as much money as you want in technology. And technology alone is not going to save you. It depends on how it's implemented. Does it fit within your your fraud program, your risk tolerance? Are you causing too much friction, not enough friction? There's all kinds of things that go into that. But if someone walks into a branch and says, I would like to withdraw twenty thousand dollars. And the teller says, okay, here you go. And doesn't ask questions. Perhaps it's a valid withdrawal, right? Obviously need a CTR there. But if someone's withdrawing a large amount of cash, maybe they do this all the time. And it's not that big of a deal. And I think I was, I don't know if it was the last time I was on here? Or a while ago? We talked about my father getting scammed. He had someone call him and talk him into withdrawing twenty thousand dollars in cash and put it in a Bitcoin ATM. And I'll save my my hatred of Bitcoin ATMs for another day. However, had the, had the teller. Or someone else at the branch said, hey, you're an 82-year-old man. You've never come in and withdrawn any cash. Why do you all of a sudden need $19,800? Perhaps he, they would have stopped that that scam from happening.
Hailey Windham
53:58
Yeah
Andrew Austin
53:59
Right? So you, absolutely, and this happens every day all across America. All across the world. Where manual procedures, or culture or, training isn't enough to effectively combat some of the scams that are going on. This this this isn't even what I'm talking about. Isn't even an AI scam. It is just a scam in general. So you know we can, we can talk all day long about AI. And how financial institutions, fintechs, so on and so forth, are vulnerable to AI. And they are. However, they're just as vulnerable to people that scam for a living. You know, we we some of the people we know. Erin West and Paul Raffile are are, or were, recently over in Nigeria.
Hailey Windham
54:48
Yeah.
Andrew Austin
54:49
Fighting scammers. And like going with the police and federal agents to like arrest scammers in person. Like these people do it as their job. They have shamans that are blessing the scam. Before they they go out and do this. So it, if if, your job and everything you built your life around, and this is what I told my dad, like you know, when when he got scammed, he felt so terrible and so stupid. And my point to him was, you know. You you worked for 30, 40 years as a as a cobalt programmer or DBA, this and that. Everyone came to you. For questions, you know. Because you were the expert. These people are the same way with scams. They know how to emotionally manipulate people. To cause that sense of urgency. To give them that call of action. Where they are actually going to do something. To take their own money and send it to a scammer. That is always gonna happen. Maybe it's not the oldest profession, maybe it's the second oldest profession. But there are, there are always gonna be people that scam. That manipulate for their own financial gain. So going back to your original question. Like those manual processes. Those people. Those frontline defenders. Are of utmost importance. And the training that we provide to them, or the financial institutions provide to them. The awareness and adherence to those procedures. Is of utmost importance.
Hailey Windham
56:18
So true, so true. Culture is everything. I'll say it till my book comes out. But culture's everything.
Andrew Austin
56:24
Eagerly anticipating that.
Hailey Windham
56:27
Yeah, yeah. So okay. We're looking ahead now, right? Then the next five years, right? Which I should probably say the next two years. The way that technology is innovating and at the speed in which it is. When you look ahead though, Andrew, what what concerns you most?
Andrew Austin
56:43
The speed at which AI technology is proliferating and getting better. It's, it’s difficult to keep up with. You know, like I said. Like I found this lip sync technology last night. I don't know how long it's been out. It's probably been out for a while. But, you know, it's you just gotta look for it. It's not hard to find. You just gotta look. And it gets better and better and better. You know, two, three years ago we're watching Will Smith and his mangled face eat spaghetti. Like I I I brought this up. It's flawless now. If it's flawless now, what more can we do in a year, two years, five years from now with AI technology? And can the defenses keep up? I certainly hope so. It's not even keep up, it's catch up. It's catch up and stay nearly parallel with them. You know, fraud trends, you're always just a step or two behind the fraudsters. And either your model or your rules, have to be updated or retrained, to keep up with the latest trends and fraud. Whether that's onboarding, transaction fraud, whatever. So can they, can they, AI tools that we're building to catch the AI, get back on that level, catch it faster. And then stay on the same pace of development as the tools that are being used to exploit those systems?
Hailey Windham
58:11
Okay, so what excites you?
Andrew Austin
58:17
What excites me. You know, I I see a lot of the the work that the fraud industry is doing with AI. And Sardine is is working a lot with with AI. Building AI agents. I think there's a lot of hesitancy within the financial institutions to fully adopt some of those. Rightfully so, right? Because regulated institutions are scared to allow AI to really decision things in most cases. Especially in compliance. Where, you know, maybe you miss a CTR, you miss a SAR. Or you know, the the wrong information is put in because the AI hallucinated. And now you're under regulatory scrutiny. And no one wants that, right? But I think as as we progress, there are going to be more use cases for AI for operational efficiency. Which does not necessarily mean like job loss. It it, the way we are building them, and the way I've I've seen them built in the industry, is is really around providing the analyst, the investigator, with additional information. To make better decisions quicker. And what that, in theory should do, is open up capacity. Maybe, maybe there are less investigators and they're more data scientists. To more fine-tune those things. And and catch additional information that's maybe slipping through the cracks now. So that's where I'm hopeful. Is that, and excited about, is that as AI tools progress on our side, that they're they're helping financial institutions catch more fraud. To catch more money laundering. To catch it faster. To provide law enforcement with additional information that's needed to investigate these. And ultimately you know, stop, prosecute, imprison criminals.
Hailey Windham
60:20
) Yeah. Put em in shackles, man. Okay, next question. Just, I'm gonna kind of combine a couple. Because they're kind of asking the same thing. Which is, you know, if you're advising a CISO today, you know, what should already be on their roadmap? Or, you know, what should financial institutions in general be investing in over the next few years to help combat this effectively?
Andrew Austin
60:45
That is a great question. The technology to combat it, I don't think really there are there are some things that we use to augment it. I don't know that it, the core foundation of the technology has changed that much. We we already talked about device intelligence, behavioral biometrics. I think those are are critical components within a fraud program. A a strong transaction monitoring program, strong fraud procedures. I don't know that I would say, hey, CISO, you need to invest in a bunch of anti-AI stuff. Right? I think awareness of it is, is a key. And then looking at your existing process and procedure to analyze could What do we need to do within our current flow to combat that? Right. I I'm yes, I want you to buy technology. I work for a software company. But I don't want you to buy technology just to buy technology. I, going back to like the the people component of it. Like the people are important. But, the business process is just as important. And understanding where you can put in friction, what type of friction that is. Is, you know, you can do that a lot of times with existing tooling. When we talk about AI and financial institutions, around the financial crime space, I think there's a ton of opportunity. For it to be used, as I as I mentioned earlier, for operational efficiency. To pull out insights. And create connections between individuals, or parties, that were previously unknown. So if, you know, a lot of, a lot of transaction monitoring programs, just pump out alert after alert after alert. Maybe they don't consolidate alerts. Maybe they don't look for related parties. But using technology, some some of its AI, some of machine learning, that can pull that information together. Into like a a case, or a package. An alert package to review. You know, maybe it's five different alerts that are pulled together because of related parties. That is where I'd be investing. Is is getting better at, what is an investigator's core competency? Investigating, right? It's not a trick question. They're they're good at investigating. They're good at, like drawing connections. And seeing things. But you know, the people always talk about, you know, I, and I and I used to do this when I worked in banking. I worked with our our financial crimes compliance team. Building a transaction monitoring platform. And I go and sit with them. And they have 20 different screens open. You've got the teller platform, the deposit platform, this platform, that platform. And they know what they're doing. Because they've done it so long. But if I can bring that into like a ever, you know. Let's let's get into industry terminology, single pane of glass. Or customer one view. You know, you you bring that. You centralize the, all the data that they need into one screen. Or as you know, consolidate, realistically consolidate, as much as possible into one screen. Give them summaries of what has happened. And show them the connection. So they don't have to click 30 times to find that Andrew Austin has a, some sort of relationship with Hailey Windham through work. Then if I can cut out that, you know, 10 minutes, 20 minutes, 30 minutes of of manual investigation, that is a huge uplift over a single day, over weeks, over months, over years.
Hailey Windham
64:16
I just think about the efficiencies that it creates, truly. Where I was the fraud fighter of one. Imagine being able to truly prioritize. I mean, I had a system. That I had a case management system, that would score things. And I would look at all of the things that were scored at, you know, they they'd put them in order. So if it was ninety-nine to eighty. I could see what was scoring scoring the highest. And, but when I would look, there's so many false positives. And there's so many things that literally, if you just looked at it, you knew it wasn't a fraud case. It was, you know, a new account opening. And they were setting up payments. Well, I can't tell you how many times I've had an account and I was like, yeah, they have this function of remote deposit. Well, if I want to open up my online account bank you know, online account, then set up remote deposit. Well, that's why I opened online banking. Was so I could have remote deposit. But I can't have remote deposit for fifteen days. Like, okay, that's,
Andrew Austin
65:16
And you, or that customer is going somewhere else.
Hailey Windham
65:20
Exactly. 100%. I'm going somewhere that will allow me to open it, and create it that same day. Because why have it, if I can't use it? And then also, I mean, I'm just gonna forget about it. And say, okay, fine, whatever. So Andrew, I I can't thank you enough for deep faking me. Making me say things that I never said before.
Andrew Austin
65:38
My pleasure.
Hailey Windham
65:41
And for coming on to Fraud Forward. Is there anything else you want to add to this conversation, that maybe I didn't ask? Or, you know, any any closing thoughts?
Andrew Austin
65:48
I think we covered most of it. I think you know, awareness is just huge, right? I do not like when when companies use these types of things, to scare clients. Right? You take a video. It's like, is this me or not?
Hailey Windham
66:04
Right.
Andrew Austin
66:05
It, does it matter? I don't know. It doesn't really matter. Like you, you're just, you're sowing doubt out there, right? Like that's why I like like shows like this, right? I can come on and have some real talk with you. And say this is what is actually out there. And and talk about, yes, it's there. Is it everywhere? Mm, kind of hard to say. You need to be aware of it. You need to look at your own processes, procedures, tooling. To see what gaps you have as a financial institution, credit union, bank, Neobank, FinTech, whoever you are. And see where you're vulnerable, to all types of fraud risk. Not just AI fraud risk. And you know, hopefully you have the expertise in-house, to to kind of close some of those gaps. And if you don't, work with consulting or vendors to try to close those.
Hailey Windham
67:03
So true. So thank you again, Andrew. I just appreciate your friendship. And you're the only person I would trust to create a deep fake of me.
Andrew Austin
67:12
Thank you.
Hailey Windham
67:13
Although no one else is really gonna ask permission. They're just gonna do it, right? So you know, I just wanna, you know, tie this episode up with a pretty little bow. And, you know, today's conversation, it wasn't really about artificial intelligence. But it was about trust. Fraud has always been built on manipulating trust. AI doesn't change that. It just gives criminals a far more convincing disguise. So if today's opening made you a little uncomfortable, good. Because that's exactly how your customers are going to feel if we don't start preparing for this now. Andrew, again, thanks so much for joining me, helping us make this conversation practical instead of sensational. And everyone listening, we'd love to hear your thoughts. Have you encountered deep fakes in your organization? Are you preparing for this now? Or is this still on the horizon? Join the conversation, let us know in the comments. As always, stay vigilant, stay informed, keep moving fraud forward. And I'm supposed to say smash that subscribe button. Thanks a lot.
Hailey Windham
68:21
Thanks for listening to Fraud Forward. Remember every conversation, every connection and every insight moves our industry one step closer to stronger fraud defenses. If today's episode sparked an idea, share it with your team or tag me on LinkedIn. I love hearing how you're moving fraud forward in your own organization. Until next time, stay curious. Stay resilient. Keep moving. Fraud forward.
View full transcript
