Preventing Predation & Encouraging Innovation in Fintech Lending
19 Pages Posted: 27 Jun 2019
Date Written: June 18, 2019
More than 20 years ago, IBM's Deep Blue vanquished chess grandmaster and reigning world chess champion, Garry Kasparov, in a pair of best-of-six matches. Since then, numerous companies have invested large sums of money to develop additional game-playing, machine-learning algorithms. They have enjoyed significant successes. For example, in 2016, AlphaGo demolished Lee Sedol, a world champion Go player, 4-1 in a series of matches by, among other things, making “a move no human would ever play, stunning experts and fans and utterly wrong-footing world champion Lee Sedol.” In addition, Libratus defeated “four of the top-ranked human Texas Hold'em players in the world over the course of 120,000 hands” during a twenty-day poker competition in 2017. Other algorithms have also defeated top players in checkers, chess, Jeopardy!, and Scrabble.
These companies are not trying to dominate your family game nights. Instead, they are betting that many of the techniques and strategies used to create world-class game-playing algorithms can be deployed for other purposes as well. In other words, they are looking to create general-purpose artificial intelligence (AI) systems. One use for AI systems is to improve credit-underwriting models. The regulation of these types of AI systems is the focus of this paper.
This new breed of lenders, often called fintech lenders, are amassing hordes of data (Big Data) and using machine- learning techniques honed in the game-playing context to improve existing credit-underwriting models. Fintech lenders commonly use nontraditional, tech-centric methods to market themselves to prospective borrowers, evaluate borrower creditworthiness, and to match prospective borrowers with sources of credit. Examples include Lenddo and ZestFinance. Fintech lenders often claim that they can lower loan origination costs and loan default rates, thus improving credit accessibility.
This essay proceeds as follows. First, it briefly explains fintech lending, focusing on its differences from traditional lending. Next, it highlights the promise and peril of fintech lending. Finally, it considers how state and federal regulators might deploy their powers to prohibit unfair acts and practices to encourage the promise and avoid the peril of fintech lending. A brief conclusion follows.
Keywords: fintech, lending, marketplace, algorithms, Big Data, machine learning, ML, predictive analytics, AI, artificial intelligence, UDAP, UDAAP, fair lending, credit, underwriting, shadow banking, AlphaGo, DeepBlue, Zest, regulation, FTC, CFPB, attorney general
Suggested Citation: Suggested Citation