Dave mentioned that the idea behind using AI to access true credit risk came from the methodology they interview fresh grads as potential employees without much work experience at Google. We can tell Google is a successful company and they have many capable people. If they are decent people as employees, why shouldn’t they be approved credit at a low rate? I am sure they don’t use the exact parameters but something similar.
Also, in a video, Paul Gu mentioned, their strategy is starting with unsecured personal loan first because it’s the most risky type of loan(AKA hardest). it’s because a person can do whatever they want with the money for any purposes and there’s no collaterals. After they are good at the unsecured personal loan, they can move to other less risky (secured) loan such as car loan, mortgage etc.
I like how they started with the most difficult problem first rather than easist problems first. Then easier problems will be piece of cake. it’s unconventional: get high hanging fruits first then get low hanging fruits. I have a feeling this idea partially came form Paul GU. It’s a very an asian philosophy: live a hard life first then easier life comes.
Without AI lending, what would be unsecured consumer lending like for people without good credit history? It’s predatory lending. It’s large number of responsible people (90%)subsidizing small number of irresponsible people. (10%). It’s like how insurance industry operates. The problem is lending is not and should not be insurance. We need precise determination of true risk and charge fair interest rate accordingly. So if personal A is most risky, charge 30% interest or even reject. If personal B is less risky, charge 8% interest. Then it’s fair. It’s capitalism not socialism.
I’ve read a news in Canada where a women borrowed $10,000 and ended paying $25000 in just a couple years! She was charged 30%+ interest rate. It’s higher than credit card rate! The question we should ask is if she can paid back not only $10000 but $25000, why wouldn’t she get a lower interest rate? Instead of charging 30% to everyone? I see Upstart is democratizing lending.
“Dave Girouard: In those early days at Google, we were hiring thousands of people each year, often recent grads with just a few years work experience. We developed an algorithmic approach to evaluating candidates and predicting whether they’d become successful Google employees. So it struck me as strange that these same individuals - known as “thin files” in the credit world - couldn’t get a $10,000 loan. Industry-standard credit models considered them risky by default. My co-founders Paul and Anna and I were inspired by this, and came to the belief that using modern data science to model credit risk could significantly improve access to credit and reduce the risk in lending.”