Leaders in Lending
Leaders in Lending

Episode 78 · 1 month ago

Fair Lending and AI: A Conversation with Upstart’s Chief Compliance Officer, Annie Delgado

ABOUT THIS EPISODE

Access to credit is the key to opportunity and upward mobility for millions of Americans, yet less than half of Americans have access to prime credit today. AI and machine learning are helping lenders identify creditworthy borrowers without increasing risk, but some have questions about how these new technologies apply to Fair Lending laws.

Annie Delgado, Chief Compliance Officer at Upstart , has been actively engaged in working with regulators on fair lending practices and shares how Upstart works to prevent bias in its AI algorithms.

Join us as Jeff and Annie discuss:

  • Policy implications of use of new technologies (like AI) in lending
  • How to interact with regulators in this new space of AI and lending
  • Why fair lending is a policy issue we need to be concerned about

You're listening to Leaders and Lending from Upstart, a podcast dedicated to helping consumer lenders grow their programs and improve their product offerings. Each week, here decision makers in the finance industry offer insights into the future of the lending industry, best practices around digital transformation, and more. Let's get into the show. Welcome to Leaders and Lending. I'm your host, Jeff Keltner. This week's episode features my conversation with Upstart Chief Compliance Officer Annie Delgado. Now, as Upstart has been at the forefront of applying AI to lending, Annie has been at the forefront of applying regulations to AI and lending and figuring out how to interact with regulators, how to think about questions like fairness and bias, how to compare an AI powered future with current status quo, and so it's a really thoughtful conversation about how to think about the problems we're trying to solve with the application of these kinds of techniques to the lending ecosystem, why that matters, what the critical questions are, and how lenders and fintech companies can engage with regulators to help drive forward the state of the industry so we can actually unlock more and better price credit for the American consumer and more safety and on is for the American financial system. So she's thought long and hard about these problems and it's a really fun chance to dive into them with her. So please enjoy this conversation with Annie Delgana and you welcome to the podcast. Thanks so much for making the time to join us. Thank you for inviting me. I'm excited to be here. I know I've been looking forward to this one. Uh we have we have lots of animated conversations that are not recorded, so I figured it'd be fun to have one on the record, if that's the right phrase is on no pressure, no pressure. But uh, I've been starting all my podcast interviews now with this question about how people got into banking. And you've been in the in the finance space longer than many of us here, and I'm started to give me a little bit of your history and how you wound up both in the financial services industry and then at some crazy fintech company. Yeah. Absolutely so, um, believe it or not, I did not, you know, go to bed as a kindergartener dreaming of becoming a compliance officer. I sort of found my way into it. UM. I did actually spend a number of years dreaming of becoming a lawyer, and um what happened was that I finished my sort of undergrad career right around the time of the beginning of the financial crisis, UM, back in two thousand seven eight. And I had the value of having friends who were a few years ahead of me in high school kind of moving home around that time having finished law school and saying, oh my goodness, we're sitting here with loads of student debt, and um, it's hard for us to actually find that first job as a lawyer right now. And so the risk minded side of me, who was always a little bit um you know, good at risk assessing and deciding what was a good choice or not for the next steps for me, decided I better work for a few years and make...

...sure that I really want to go to law school if I'm going to do that, versus taking some other path in life. And so I ended up, um, you know, because the labor market then, of course turned a little bit negative. I ended up casting an extremely wide net in terms of what I would do for my um first job, and I ended up getting a job UM as a QA analyst at a mortgage company, and I was, let me tell you, a really exciting time to work in mortgage. Right around that time, UM has not only had the meltdown sort of just been involving right in front of my eyes, but also we started to get a lot of new regulation coming out with GF the CFPB got formed, Dot Frank happened, UM, and so it was sort of this extremely exciting space and I found, um, you know, a niche for myself in terms of passion for public policy, particularly as it relates to UM financial services, and ended up deciding to to stay in financial services and really to stay in the complying space within financial services. And so UM, that's that's why I built my career in that I love it. And you were like the fourth person and tell me they were going to become a lawyer and then ended up in the banking space. Maybe I I sent a trend coming. I don't know what it is. It's the it's the better alternative back up to something closely related, I guess. Um. Yeah, So you know, the thing I think that makes the most sense for us to talk about is kind of the regulatory landscape, AI and lending. You know, there's lots of topics there, um, but the one I wanted to start with first was kind of thinking through why it matters, Like to me, there's this question of, um, how do we get the regulators comfortable? What are the written measurements? Yeah? Yeah, yeah, but this of course, like should you be doing it in the first place? And what problem are you trying to solve by throwing AI at it? And so I kind of wanted to start there because I feel like without that, there's kind of like, why are we going through the rest of this if there's not a problem to solve? Um, So what is the problem to solve? And how do you see the upside of applying something like like what's the reason to do that? Yeah? Absolutely, um, And I think that is the right the right question to ask anytime you're trying to change the status quo or innovate, Um, you better make sure you're actually doing that for a reason. And if you have a really finely tuned system that's working really well, there's no real reason to do that. But um, I would argue, and I think it's uh, maybe not discussed enough that our system today has way too much inefficiency and way too many people are left out of the system. UM. And I think that it's just something that's not discussed publicly enough, the sort of breakdown in terms of access to credit that we have, UM, you know in the United States, where we have a large population of people who would be really good bank customers bank borrowers, who are instead declined for a loan because banks aren't exactly sure how to underwrite the risk associated with with that particular application. And so UP started a study on this a few years back. We found that something like of people UM have access to what we would call prime credit, but eight percent of people have never defaulted on a own And that tells you that there is a huge opportunity in terms of...

...people who should be qualifying for loans who are not qualifying for loans. And then there's an even um sort of bigger population who aren't even measured by that study, who are what we call the credit invisibles UM, who either maybe don't have a credit score, the recent immigrants, they're just building their credit history, etcetera. UM. And so the opportunity to sort of have a more efficient system that's more inclusive for people is vast and it really needs to be solved because most of us are not in a situation where we just have unlimited funds that our fingertips whenever we need them. We actually need to borrow money for things like cars to drive to work, or to go to to go to college, or to go to law school if we want to. Most of us don't have the money sitting around to do that. We have to borrow it. Um. And so if you have a system or only half the people are able to borrow it, that's a broken system because everybody should be able to have access to that money so that they can achieve their dreams. Um. And that's really the problem that we're trying to solve by unlocking some these innovations like AI in lending. Yeah, I think it's such a such a critical point. And you know, sometimes I think people wonder if credit really matters, if we should make credit more accessible or not, And I think it's so easy to get lost in in that, you know, that question versus the reality that like access to opportunity and mobility is really in many cases for many people gated by access to credit. And so credit is not just credit, it is it's your ability that to participate fully in in kind of the American dream and the opportunity. Yeah, that's exactly right. It's the path to homeownership, it's the path to employment for a lot of people. It's the path to sometimes paying for, you know, the healthcare that you need so that you can stay healthy to do those things and provide for your family. It's the it's the access to going to college. It's it's it's really beyond just credit. It's really building your future. Um and that's why it's a problem we've got to solve. So one question maybe that comes up, and you may not have a perfect answer here, but I'm just curious your thoughts is, how has this not been a problem? Like why is it a problem that's not addressed, that's spoken quietly, that's not something we focus on. And you know, it's interesting to me that it's like, as a as a math kind of person, I look at that inefficiency that's a huge problem, Yet for many of the financial issue doesn't seem like a huge problem. I'm curious why you think what that disconnect is between that opportunity you see and and the kind of sense of there's a problem here and something we're talking about that that doesn't feel as urgent to me in the industry at large as it does to maybe you or I, UM, well, I don't have a perfect answer, that's true. I think that hard problems are hard to talk about, and so a lot of times when you have things that are really hard problems, people would prefer not to talk about them. And I think that there is sometimes a disconnect between who the problem is is facing and you know, the sort of business needs on the other end. So in the banking industry, if you aren't lending to everybody, but the people you are lending to are paying back and your business model is sound, etcetera, you maybe don't have as big give a concern about...

...the people that you haven't been lending to. UM. But as a society, we should be really concerned about that, right because there's you know, some of the problems that we just sort of we're talking about in terms of you know what those what the people left out of the system are are truly getting left out of UM and so I think there's sometimes just a um a disconnect there. And again I think the underlying problem, like I said, is it's a hard problem it's a hard problem to solve, it's a hard problem to talk about, UM, And it's just something that that, in my opinion, should be brought more to the forefront of conversations when you say, at the societal level, that feels like something government, regulators, policymakers should care about, right, Like I can totally understand if you're a lender and you go, hey, I make the loans I need. From my business point of view, the volume is right, the risk tolerance is right, that the losses are right. I don't know my business is working as it should um. But maybe at a government or a policy we've got people who aren't accessing credit. That's a problem. Have you found We can talk about kind of other topics of engagement later, but I'm I'm curious if you see the focus on that from regulators, a sense of like, hey, we do have a need in terms of access to and and of course the other side of the coin from access to is the cost of credit, because both of those are really related to in accuracy in the system. Is that something you see regulators policymakers interested in focus on is something they want to see their influence and kind of direction solved within the industry. UM. I definitely think that there is a focus on it and there's conversation around it. UM. From a policy standpoint, it's much harder problem to solve because you don't need to only be worried about the access you also need to be worried about the consumer protection aspect of making sure right. You can easily solve access to credit by giving everybody alone, but then you've caused this other problem of putting people in loans that they can afford right, which is a predatory issue. And so I think it's just an incredibly complex topic, and so the regulators and policymakers have to be looking at it from all of those different and angles and sort of striking the balance and and having all those conversations simultaneously, which frankly is what you know, UM fin techs and bankers have to do as well. All of us have to be sort of simultaneously worried about all all sides of that of that problem. So you're not an AI expert, well, you're somewhat of an a expert, But like I've talked to you know, Paul and others in the space about how the how it works so I'm gonna for the purposes of this conversation, let's take it as a given that some form of advanced analytics can help bridge that gap between that you know less than half and that that you're talking about. I don't think we need to dive into how that's done and the results we've seen there. But assuming that's true, assuming you know, some sort of advanced analytics machine learning can solve some of this problem you're talking about, what are the concerns other than raised about is it being done fairly equitably within the state of regulation, Like, give me a sense of the landscape of concerns. Someone says, hey, I've got a better mousetrap, it's actually is doing the job. Um, one are the things I have to be thinking about, concerned about measuring,...

...monitoring to make sure I'm doing it in a way that's within the bounds of the law. Yeah, um, absolutely well, And as you mentioned, UM, I am not an AI person. I'm a you know, financial services person and a legal regulatory person. Um. But you know, having been at up Start for the last I think about it all the time, and um, really, the sort of way I think about this is that if you think about lending. Everybody should be concerned about bias and lending, no matter what system you're using. Whether it's a human underwriter or a scorecard or a checklist, or logistic regression or an AI model, no matter what the system is, you should be concerned about bias and lending. And so when I think about this problem, I sort of break it down into there's three components of any sort of lending assessment. There's what goes into the assessment, there's the assessment itself, and then there's what comes out of the assessment. And so, UM, let me just sort of break down what I mean by that, UM and sort of where why there is risk and lending and fairness risk and lending. Right now, there's a sort of suite of what all called traditional variables that everybody uses to make lending decisions. UM, things like a credit score, things like a person's income, their debt to income ratio, etcetera. Everybody uses those to make their assessment. Those variables in and of themselves are not pristine from a fairness standpoint. When you think about something like we know that women on average are paid less than men um for the same the same job. If you're using their income, that's going to be a different assessment that's going in UM, you know, on a protected class basis. Same things if you look at something like a credit score distribution, we know that UM in the full sort of spectrum of credit scores, the lower end of the spectrum tends to be more heavily populated with minority groups relative to the higher end of the spectrum. So if you're using that as one of your only inputs into the system, you're introducing the potential for bias, right, And so you really need to understand what that what that sort of effect is. And then the second part of the puzzle is the actual decision making itself. UM using a human example of a human underwriter, there's all kinds of assessments made in a humans brain and connections and previous biases and unconscious biases, etcetera that are that are going into that assessment process UM that you don't always understand. But what you can understand from a compliance standpoint is the output, right testing the output was this person consistent and how they made their decision, did similarly situated people get the same result, etcetera. And so really the beauty of an AI system is that rather than having a hundred different human underwriter brains that you don't understand, you have one computerized brain that you can literally pull the curtains back and understand. You can look at what the assessments are. You can change the assessments if you don't like them.

This brain itself can be trained to not use hard cutoffs, like you know, credit score above X is good and below X is bad. This system can really understand more nuanced things relative to something like a checklist sort of model. And so that's why I believe that using AI and lending is actually the solution for these biases, as opposed to it's not a new risk that's being introduced. The risk of bias and lending has been around as long as people have been lending because of societal issues, because of you know, human brain power issues, etcetera. And I really think AI is actually the solution. UM. And so the second part of your question is what to measure And the answer to that is everything UM in my opinion, right, And that's the UH, that's the compliance officer in me. Right. You have when you can have all of the data and you can measure the data, and you can can look at what is the actual impact that this model has had. What decisions did it make, who did it approve? Who did a decline at what price? You know, all of those types of things. Measuring those things on a consistent and ongoing basis is so critical to understanding how that computerized brain is working so that you can constantly be tuning it to you know, advance your mission of reducing disparities and lending and improving access to credit in lending UM. And that's really, to me the crux of what we're trying to do at upstart M. There's there's a lot in there. It feels like in that analysis the similarly situated person is doing a lot of work in terms of how you think about and analyzing. How do you think about that? I mean, like that's in a traditional model, how do you define what a similarly situated person is? That credit score oriented? That's UM that phrases a lot. Are you treating similarly situated people the same assumes some sort of definition of similarly situated That it makes sense, yeah, I mean, and that that is the challenge. I think different institutions have different definitions of how they how they look at that, but Um, you know, what we see constantly in the data at upstart using the sort of more variables in using AI is that most people are not similarly situated to other people, right. Everybody has a unique set of circumstances, that unique background, and what's true for me might not be true for you. And so having a really customized, personalized, tailored approach to understanding an individual's risk is the key to sort of improving the accuracy of the model so that more people are approved for credit as opposed to again just cut up, cut out because of some uh you know, hard hard cut off what I mean, traditional lenders when I talked to it, and I'm always kind of um hearing this approach. Talk about you know, compensating factors where that human judgment comes in, so that most of them when I say this thing, well, I we don't use that hard cut off. Really there's like the hard coop and then there's the ability for the human to make judgment about someone where there's some reason to think that a low credit score or a high debt income is not you...

...know, such an issue in this case. Talk about how you think about humans looking at other compensating factors in a credit file versus what an AI model is doing. So it feels like, you know, that would be the answer to Hey, you're talking about hard cut offs is the problem, but like we have ways to get around that too in a non AI driven system. Why do we need AI so kind of compared and coddress how you think about that concept versus what's going on in the system like up starts. Yeah, absolutely, I think um. For me, the big difference is compensating factor versus statics statistically rigorous compensating factor, because when you introduce compensating factors into an underwriting decision, what you're really introducing is human discretion, and human discretion is fraught with the potential for bias, and so there's a lot of concern for me in terms of, you know, who asked for who? Who knew to ask for an underwriter to look at a compensating factor that's gonna, you know, potentially have differences across different protected classes. Who did the underwriter actually grant that too versus not what do they think counts as a compensating factor based on their own set of experiences. There's just a lot of sort of room for judgment calls and discretion and that process that UM introduces a whole new sort of suite of fair lending risk in my opinion. Whereas again, if you have it done by an a an AI model, that is, you know, making a statistical assessment based on actual real data and mathematical approaches, then what you get is a consistently applied compensating factor where the model recognizes, okay, this thing of the time is a good signal. So across every single application that I see with that particular set of circumstances, I'm going to see that as a good signal as opposed to it being sort of applied one off based on is somebody knowing to ask for it or some other person knowing to look for it? All right, Annie, I wanted to ask you about kind of fair lending laws because I think this is a really important a area of regulation and consumer protection. And you know, all lenders are subject to a variety of fair learning regulation. But as underwriting shifts both to become automated UM and then maybe more interestingly, to become more sophisticated in the types of data it uses or the analytical techniques. UM, do you think those fair lending laws are kind of up to the task or the revisions that are needed. How do you think about the state of those laws as as the state of how we do underwriting changes. Yeah, I think it's a good question. I think that the truth about the laws is they're very well written and there's not a ton in them that probably a lot of people would disagree with from a conceptual standpoint, myself included. Um, They're they're written intended to achieve exactly what what we want, which is, you know, not to have discrimination and credit decisions and to avoid some of the um easy pitfalls. Where where that can be introduced, I think where there is a lot of opportunity for advancement is clarification of how those laws can apply to existing technologies...

...and newer technologies, because there's a lot of concepts within those frameworks that reasonable people can disagree about. So when you think about something like a proxy, right, we can see through data that all variables are sort of correlated to other variables, and whether that is a strong correlation or loose correlation really sort of changes. What is the level of correlation that makes something a proxy for a protected class versus versus not? Is something that you know reasonable people can disagree on. I think also the application of how we do testing for fairness and UM what outcomes are good outcomes versus bad outcomes is another really interesting thing that UM you know, needs some clarification right now. They're sort of existing framework where people have three parts of their of their analysis that they do. The first is to look and see whether there is evidence of differences in treatment amongst protected class. The second pillar would be to look for um, you know, business justification of why that might exist, is it to meet a legitimate business need? And then the third is for less discriminatory alternatives and UM the breakdown that you see and how this is applied is that UM people have different approaches for how deep they're going to dig on something like confirming legitimate business need or confirming that there is no less discriminatory alternative. What's a reasonable amount of um, you know, research and commitment and ongoing sort of model redevelopment that is suitable in those sort of searches. And so because those two pillars of legitimate business need and less discriminatory alternative are very challenging to solve, and there's if there's disagreement about what the right metrics are a lot of people in instead try to just manage to that upfront UM adverse impact ratio, the ai R metric that a lot of people use. And the problem there is, we get we see just ongoing in the mar get some weird behaviors related to managing to that metrics since the other two are ill defined, um things like you know, limiting who you're marketing to or that type of stuff. And so when really the purpose of of the spirit of the law is to you know, get people access to credit that they need in fair and affordable ways, a lot of times the way that those laws are actually applied in institutions is almost counter to that policy objective because people don't know exactly how to achieve that in the sort of existing framework. UM. So, one of the things that I'm actually super excited about. Upstart recently launched and announced a new coalition which is really designed to sort of tackle this problem of bringing together um, you know, a lot of different voices to the table to share transparently, um best practices and thoughts about some of these these difficult issues. UM. You know, I mentioned before on some of these things, reasonable minds can disagree, but everybody's got a really valid perspective and point of view here.

So bringing together industry, regulators, you know, special interest groups, etcetera to the conversation to talk about what are the issues, how do we resolve for them? What are best practices? UM I think is a really important and valuable opportunity that can really move the needle in terms of advancing some of these conversations and hopefully providing you know, a framework for UM the broader uh, you know, industry ecosystem as well as regulators and policymakers about how we think about these types of things in the future. That's the more than fair is what you're talking about, right, that's right, more than fair. Where do people find more information about more than fair? Well? So, actually, we just launched our our website more than fair dot com and um we also have a blog post up that that you can read more about it as well also, and I think the name just begs the question what does it mean to be more than fair? Why is fair? Maybe an incift? I mean that seems to apply that fair is not of sufficient maybe a useful test um, but not a sufficient test for if we as an industry are doing all that we should to help disadvantage to marginalized traditional underserved communities. What is it that the coalition thinks is important that's more than being fair? Yeah, I think, Um, I think the primary problem is the sort of Um, if you're just talking about fairness or parody, what you don't talk about is inclusion or access and so uh, that is a big problem in the ecosystem today. People can achieve parity in terms of outcomes and decisions amongst two groups by saying I'm only going to lend to the prime end of the spectrum and I'm only going to market to them, So everybody that I market too is approved, and um, you know, there's there's parity in my outcomes. But what happens when you do that is the subprime end of the spectrum, which ends up because of our sort of societal and equities being heavily towards immigrants and um, people of color, those are the people that are then left out by that system. So on its surface, it might look fair, but it's actually not inclusive and it's it's again not solving the sort of actual societal need. Um that we have to to really revolutionize how we think about getting people access to credit so that when somebody needs, you know, a small dollar loan to change their tires so that they can get to work, or they need a mortgage, or they need an auto loan, that they have the access available to them at the time that they need it so that they can keep their life moving. Mhm. That's interesting. Yeah, I think it's it's not always intuitive to people that the idea of fairness can can run counter to the idea of inclusion. Um, but I think that that's a really interesting element. I think that's the same one you were talking about in terms of how testing for fair lending and how we need to think about approaching this, because I think there you know, is you and I've talked to the past, there can be penalties for people trying to be inclusive in terms of the measurements we might think of on a fair lending point of view, and that seems to be not the intention of the law, but the reality of the application of the law. What you're talking...

...about clarifying. Yeah, absolutely, I agree with that. Good I like when people agree with me. Is unmust to learn something in the conversation so I can speak intelligently about topic. Awesome well, that's all the stuff we wanted to cover today. But I do have three final questions that I ask every lucky participant as a guest on the podcast. So I'm gonna throw a match now, UM and then and then you can get on with your day. The first one is what's the best piece of career advice you've ever got? UM? Yeah, so UM. I actually tell this story to a lot of people that I interview and a lot of people that work for me, and it's the story of UM one of my first interviews when I was starting out my career in financial services, I sat down with the interviewer and I asked what I thought was the appropriate question, which is, if I get this job, what is the next step for me? If I if I, you know, grow and UM. The response I got was completely not one that I expected. UM. The some sort of sat back in their chair and they said, Hey, if I have to hold your hand, this isn't the right job for you. You're gonna just go as far as your own ambition carries you. And I think it was really good advice, especially for UM, you know, a young woman starting out in financial services, which is primarily UM an industry dominated by um older men, or at least older than I was at the time, men um of just being willing to raise your hand and sign up for more things and and believe in yourself and say, hey, let me try that, let me show that I can be capable of learning that thing. Um. And so getting this sort of runaway from a from a mentor that you know, just just show up and do what you solve the problems you think need to be solved, I think was really good career advice. I love that. And and empowering people with the confidence to follow their own ambitions and intelligence and thoughts and ideas and to put those four I think is so important because it's it can be tough to do, particularly early in your career, to have that confidence. I'm just gonna like, I disagree with the loss because I think they're wrong and I know more that's always comfortable, even even for today, I still find that so good advice for everybody. Second question, what's the best advice you've gotten about the consumer banking or consumer lending space? So I'm gonna tell you like a silly little story, and really it's a story about status quo bias um, which I think is a big problem in UM, in consumer lending and in financial services generally. So this also came from a former colleague that I am used to work with, and she would always say, don't cut the ends off the roast. And I would say, what are you talking about? What does that even mean? I have? You know, how does this apply to my life? And she would tell this story about, you know, some person goes to you know, cook a roast and they cut the ends off and their child asks why do we do that? UM? And the person says, well, my mom taught me to do it this way. So they go to the grandmother and say why do you do that? And she says, well, I don't know, my mom taught me to do it this way. So they go to the grandmother and say why do we do this? And the grandmother said, I don't know why you do it. I did it because my pan was too small. And I think that that's UM. A lot of times what happens in financial...

...services is there's things we've been doing for a long time, and so we have the status quo bias that seeped in where we think that just because it's been done that way for a long time it's good, or it's right, or it's the best, UM, And maybe that's not true anymore, right, And so being willing to ask that question a hundred times, is this still the right thing? Is it the best thing? Is there some alternative approach that we can be taking to this problem that would solve solve the problem in a better way? Um? I think is something that we all need to be doing. UM, especially in an industry as important as banking. UM. And important I mean to the to the consumers that we all serve. I think it's a fascinating story and so true and always so. One of the advice I got that I think is related is, you know, be willing to ask a stupid question like probably nobody wants to ask, why do we cut the inside the rest? Because you can go like you don't understand, like you're gonna feel like you're the one person roof it doesn't know. And the number of times I've asked a stupid question that I get an email I'm so glad you asked that I to know. Either I thought I thought I was the only one who didn't know, and I was out of the out of the loop. And it turns out like there's no answer to the question, or we don't know and nobody's thought about it. It happened so often, so it's great. Um, you know, don't cut the ins off the at least ask why we're at least Yeah, And that's I mean. I say this to my stuff all the time, you know, my own stuff. I say, please, don't ever do anything that you don't know why you're doing it. If you don't know why you're doing it, stop and ask me why, because you're gonna just do it better if you understand that context, or you're gonna come back to me and say I have a better solution to solve that problem. Um. And so that's that's the advice I always give to my my staff today. I'd say, I like that just as much, and I could. I could. It's not as memorable as don't cut the ins off the roast, but maybe more immediately applicable because I don't I don't look a lot of roasts myself. Um. Alright, My last question one bold prediction for the future. What do you have for me? Yeah? And I think my bold prediction for the future is, you know, I'm a huge history um buff as you probably about me, since we've been friends for a long time. But as we look through history, we see just time and time examples of things where we had some big advancement in in society or culture, and always in that time there's naysayers who think it's too risky or uh, you know, we shouldn't be doing it, it's not the right move um. And so I think in some ways the same thing sort of is happening a little bit now with AI. There's a lot of people who are afraid of UM, of machine learning and about artificial intelligence and the risks that it posed for the future. And I think if you over index that looking at the risks, you don't also look at the benefits. And so what really should be happening is a confronation conversation about both how do we harness the benefits while mitigating the risks. So my bold prediction is that AI in machine learning and underwriting is going to help us solve some of these societal and equities that we have, because I think starting with getting people access to credit is a really important starting point and solving some of the the broader um sort of inequality problems that we have. I...

...think that we have a potential for solution in AI and mL if people really adopted and so um, that's that's sort of what I'm hoping that we see materialize over the next short short period of time here, because I think we have the technology and now the rest of us needs to just catch up to being coomfortable using it and allowing it to help solve the problems again while appropriately mitigating any associated risk with it. I like that bold prediction is both positive for the industry but also for the the American public and the consumer in terms of enabling access. Well, thank you for taking the time today and I appreciate your thoughts. And for those interested in learning more about more than Fair, you can find us at more than fair dot com, So go check it out and you can read off more on these topics. Up Start partners with banks and credit unions to help grow their consumer loan portfolios and deliver a modern, all digital lending experience. As the average sumer becomes more digitally savvy, it only makes sense that their bank does too. Upstarts AI lending platform uses sophisticated machine learning models to more accurately identify risk and approve more applicants than traditional credit models. With fraud rates near zero, up starts all digital experience reduces manual processing for banks and offers a simple and convenient experience for consumers. Whether you're looking to grow and enhance your existing personal and auto lenning programs or you're just getting started, upstart can help. Upstart offers an end to end solution that can help you find more credit worthy borrowers within your risk profile. With all digital underwriting, onboarding, loan closing, and servicing, It's all possible with Upstart in your corner. Learn more about finding new borrowers, enhancing your credit decision ing process, and growing your business by visiting upstart dot com Slash four dash Banks. That's upstart dot com slash four dash Banks. You been listening to Leaders and Lending from Upstart, make sure you never miss an episode. Subscribe to Leaders and Lending in your favorite podcast player using Apple Podcasts. Leave us a quick rating by tapping the number of stars you think the show deserves. Thanks for listening, until next time.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (86)