The Case for Artificial Intelligence in the Financial Services Industry
Introduction
I rank Elon Musk right up there with Thomas Edison, Nikola Tesla, and Alexander Graham Bell on the
brilliant inventor scale. However, I do not agree with Musk’s dire outlook on the impact of artificial
intelligence on mankind. I do not believe our world will being taken over by Killer Robots and humans
will be enslaved by them. There may be some bumps along the road. However, I envision a future
world where humans and artificial intelligence will peaceably co-exist for the benefit of all mankind.
Back to the Future
I have been working in the financial services industry for many years. I witnessed the dramatic stock
market crash in 1987, the Asian Contagion of 1997, the 911 Terrorist attack in 2001, and the Great
Recession of 2008. Humans survived each of those debacles. Now I am bombarded with gloom and
doom scenarios about artificial intelligence (“AI”).
AI is nothing new. It has been with us for decades. Most of the algorithms used in AI, or more
specifically machine learning, have been around for years. For example, I applied the Principal
Components Analysis (“PCA”) algorithm to historical yield curve movements to derive scenarios to use
for Value-at-Risk calculations almost twenty years ago. PCA is now known as an unsupervised, machine
learning algorithm. The PCA algorithm itself has not changed, only the semantics surrounding its
description.
Multiple linear regression analysis is typically used by banks to predict earnings and capital under
various macro-economic and market scenarios for CCAR Stress Testing. Regression analysis is a long-
time staple of statistical analysis. It is now known as a supervised, machine learning algorithm.
Independent variables are now called features and the dependent variable is called a response variable.
Same algorithm, just new terminology to describe it.
I bring all of this up to make the point that if civilization was going to end because of AI, it would have
already ended. At the very least, we would have already seen signs of our impending doom. There are
very few “new” algorithms used in the AI world. Many of the same algorithms that have been used for
decades are just being applied to larger datasets, i.e., big data. AI has been driven to a large extent by
cheaper data storage space and more powerful computers; not by new innovations in algorithms.
AI Makes Our Lives Easier, More Productive and Interesting
I cannot imagine how long it would take to derive regression equations for CCAR stress testing based on
realms of data without the use of software and hardware. I love math and statistics, but I have no
problem letting a computer do the heavy lifting for me. Computers can efficiently use big data and
perform mathematical calculations, while freeing up humans to analyze and derive insights from the
output. AI and humans can accomplish much more together than either can on its own.
While I do not see AI leading to humans being subservient to computers or robots, I do see a potential
impact on jobs in the financial services industry. The automation of cognitive tasks may indeed make
some jobs redundant. Smart chatbots that leverage natural language processing (“NLP”) may replace
customer service reps. Loan processor positions may be reduced due to robotic process automation.
There could also be increased unemployment among high-skill workers due to the use of machine
learning in banks. The key to staying ahead for these professionals is to brush up on their mathematics
and computer programming skills. This does not necessarily mean going back to a traditional school.
There are many free and affordable training options on web sites, such as Udemy and Coursera.
Machine learning typically leads to the automation of limited tasks as opposed to the automation of
entire jobs. AI can do the boring, mundane, repetitive, or tedious cognitive tasks, while the skilled
worker can be freed up to do more interesting and stimulating tasks. Therefore, AI has the potential to
make the work life of skilled employees more fulfilling. The employees who wish to thrive in this new AI
world will need to have a good understanding of what AI can and cannot do and where they fit into the
equation.
Conclusion
I believe that the benefits from AI outweigh its risks to society. I expect more banks to embrace AI in the
future. Machine learning will continue to be used by banks in the areas of financial prediction, fraud
detection, credit scoring, and trading. Chatbots will be used by more banks to interface with customers.
AI may have a negative impact on the employment of some workers. On the other hand, existing
workers who embrace AI and upgrade their skill-set accordingly may flourish in the new AI future.
December 7, 2019
