Loans have become an important part of our everyday life and all sorts of different people apply for them, whether it’s to invest in a business, buy a home or book a holiday they have been dreaming about for a while. In order for a loan to be granted, one should conform to a variety of parameters. The more variable applicants comply with, the better their creditworthiness is certified. Thanks to artificial intelligence (AI) algorithms are replacing humans, when it comes to analysing their credit score, as they can provide more accurate estimates.
Smart Finance is an AI application that uses algorithms to grant millions of small loans in just a few seconds of analysis.
The applicants should allow full access to Smart Finance in order to collect some of their mobiles’ data, instead of manually entering the amount of money they earn and other parameters, like those which are traditionally provided to assess the probability of the loan being repaid.
How much battery do you have in your mobile?
The data is analysed through deep learning, a method of more sophisticated machine learning that allows artificial intelligence to learn by itself.
For Smart Finance, the important data is the one that does not reveal anything important to a human being. It studies how much battery is left in the mobile phone, the speed at which the users enter their date of birth, how often they order take-away food, if they took enough time to read a user agreement and other parameters that end up forming a kind of digital fingerprint capable of predicting whether the borrower will pay the requested loan.
Smart Finance builds a credit rating system based on 1,200 data points related to user behaviour and subsequently connects potential borrowers with lenders.
One of the problems with machine learning is that the algorithm may have detected correlations in which the causes are not visible. For example, algorithms may have found a correlation stating that users who have a battery below 12% only return their loans 43% of the time.
This strange correlation has been identified by the algorithms after diving in large amounts of data, but humans are unable to process such amounts of information and are also unable to explain what the underlying relationship is. We may try to verify whether this correlation exists and is consistent, but we can’t discover the reason why a person with that percentage of battery is likely to breach his contract.
In 2017, Smart Finance granted more than two million loans per month with very low default rates, far exceeding assessments made by traditional banks. The company’s founder Ke Jiao describes the metrics used as the "new standard of beauty."
Privacy is another thorny issue raised, however, as the application dives into unsuspected limits in our daily lives data. In fact, many platforms that track the use of smartphones have access to data such as location services, telephone contact lists and call logs that can be used to later track and harass delinquent borrowers.