Current credit-scoring technology analyses about 50 data points - much more than any human could economically. Zest, brainchild of former Google CIO Douglas Merrill, analyses tens of thousands of data points at a time with blistering speed. They claim that it takes less than ten seconds to produce a machine-learning assisted credit rating on someone with zero credit history.
To train their A.I. & machine learning system they've partnered up with Baidu in China. Only 20% of the population in China has any known credit history. By discovering patterns in the data they gained from millions of transactions, they grew the approval rate of their lending business by 150% in two months.
To test for bias, Zest relies again on machine learning, by testing the system's results using the same system. It's just another step in the battle between human intuition and empirical machine-learned data - which one will prevail?
The platform was fine-tuned based on the experience Zest had working with the search engine Baidu in China, where only 20 percent of the population has any known credit history. Studying 21 different factors such as how people search and the way they traverse between Web pages, Zest discovered patterns in Baidu’s data that could be used to decide whether to make small loans to those customers for purchases like clothing. Among the things Zest evaluated was how well a person’s self-reported income matched up against their “modeled income,” what Zest calculates that person actually earned based on other behavior.