In today’s post, we’re going to take a look at how artificial intelligence (AI) is playing an increasingly important role in the rapidly evolving world of financial technology, or fintech.


Let’s start with Investopedia’s definition for fintech:

“Any technological innovation in the financial sector, including innovations in financial literacy and education, retail banking, investment and even crypto-currencies like Bitcoin.

According to EY’s Fintech Adoption Index, 33% of consumers worldwide use a minimum of two fintech services. China, by far, is the country with the highest fintech adoption rate at 69%.  By contrast, the U.S. is at a 33% adoption rate.

The Fintech Adoption Index also outlined several of the key trends worth noting:

  • Fintech has achieved the key milestone of “early majority” adoption. Much of the growth is driven by increasing awareness of fintech services. Today, 84% of global consumers are aware of fintech, up from 62% at the end of 2015.
  • The most popular fintech category is money transfer and payments, used by more than 50% of consumers worldwide. Insurance services is the category experiencing the most significant growth, today at 24% adoption.
  • 13% of consumers are considered “fintech super users,” adopting 5+ services. Not surprisingly, the age segments with the broadest fintech adoption are 25-34 (48%) and 35-44 (41%)


It is estimated that 20% of the citizens in developed countries will use AI in some capacity by 2020.

Let’s take a moment to define AI:

“Computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making and translation between languages.”

With regard to fintech, some of the most common uses of AI include:

  • Generating insights that can accurately predict customer behavior. For example, AI can learn a customer’s past behavior and make accurate recommendations about the customer’s creditworthiness.
  • Neural networks are computer systems modeled on the human brain and nervous system. These networks can visually identify customers and related documents, dramatically streamlining functions like account creation, and loan and insurance origination. AI can visually verify if documents are authentic and whether a customer applying for a loan is the person he or she claims to be.
  • Generative Adversarial Networks (GANs) are a class of artificial intelligence algorithms used in unsupervised machine learning, implemented by a system of two neural networks contesting with each other in a zero-sum game framework. GANs will be helpful in detecting fraudulent behavior, suspicious transactions and early detection & prevention of cyber-security threats.
  • AI will power the next generation of customer service chatbots, allowing them to intelligently answer customer questions and thus reduce the workload to customer service departments.
  • AI can cut processing time in half. This will dramatically reduce workloads for tasks such as processing receipts.
  • By quickly validating and double-checking data, AI can improve performance for tasks that often suffer the most from human error, such as identifying duplicate expenses or invoices.
  • By understanding “approval workflow,” AI allows companies to modify and automate myriad tasks, such as the expense tracking process.


While in the long run there are clear benefits to merging AI into Fintech, we must also be cognizant of the short-term roadblocks. The most notable of these roadblocks is bias.

The conventional wisdom is that A.I. would actually reduce bias in decision-making.  It’s generally assumed that using data to make complex decisions would seem to take people’s prejudices out of the equation.

But that’s not necessarily the case.  AI, like any new technology, is only as good as those who created it.  Which means it may reflect the biases of its creators. As highlighted in a recent Fast Company article, the bias issue is compounded by the fact that “recent advances in the technology–deep learning, reinforcement learning, and artificial neural networks–are such that even its designers struggle to trace the logic of how AI knows what it knows.”

As regards Fintech, the primary concern today is that AI, in essence, renders “black box” decisions.  Let’s say you apply for a loan and your application is denied. Current financial services guidelines require the creditor to provide you with reason codes to explain the denial. But in the world of machine learning, your application data goes into the system and a yes/no decision emerges without explanation. Today, consumers are protected against discrimination when applying for a loan. In the world of AI, they cannot be guaranteed similar protections.

Clearly, the key to broader adoption of AI into fintech is greater transparency.  A recent blog post on the White& summarized the issue this way:

“While humans may be able to identify and prevent this type of biased outcome, smart algorithms, unless they are programmed to account for the unique characteristics of data inputs, may not. To avoid the risk of propagating decisions that disparately impact certain classes of individuals, lenders must incorporate visualization tools that empower them to understand which concepts an algorithm has learned and how they are influencing decisions and outcomes.”