Episode summary: Although machine learning in finance is far from new, it is merely at the cusp of a much wider set of applications (in all segments of finance, from insurance to bookkeeping and beyond). Already machine learning has overhauled so many aspects of the financial landscape, from accounting to trading, and it is destined to have more and more impact as it develops further. Guest Alexander Fleiss and his team at Rebellion Research are developing and using AI which uses quantitative analysis to pick investments. Fleiss discusses the current status of machine learning in the world of finance as well as lesser-known niche applications that don’t make headlines – but do make a big impact on how businesses are run. He then goes on to explore the effects of future innovative applications of AI in the financial domain.
Expertise: Alexander Fleiss: CEO, Chairman and Co-founder of Rebellion Research, Fleiss has a BA in Political Science and Art History from Amherst College. His fields of expertise include applied mathematics and Hedge Funding.
Brief recognition: In 2007, Fleiss and three associates, Jeremy Newton, Jonathan Sturges, Spencer Greenberg founded Rebellion Research. Rebellion Research is a Registered Investment Advisor. The company uses machine learning technology to evaluate the value, growth and momentum of stock. Prior to Rebellion, Fleiss worked as Principal at NYC-based hedge fund KMF Partners, and as an equity analyst at Neuberger Berman.
Current Affiliations: CEO, Chairman, and co-founder of Rebellion Research, Board member with The New York Food Truck Association (NYFTA).
“If we’re honest, machine learning is synonymous with job loss.”
Mr. Fleiss has some extremely strong opinions about the job loss implications of machine learning. In his opinion, the top four US banks will be among the most “hard hit” public companies in terms of firings due to job automation. Fleiss is of the belief that other companies similarly reliance on outdated back-end processes and paperwork will likely be hit just as hard.
Fleiss speaks bluntly on the subject: “Every convo about ML ends with disrupting our economies in some way. What are efficiencies? Cost savings and cost savings, really.” He states that he worries often about people in paper-pushing positions who may find themselves without a job in the coming five years ahead.
Among the finance jobs most heavily hit by job loss thanks to machine learning, Fleiss lists the following:
- Brokers (who he believes are in the worst position)
- Bankers / tellers
- Financial advisors
- Back-office workers who handle data entry and data management
- Anyone who’s primary role relates to the handling of paperwork
In many respects, Alexander seems to be of the belief that we should in fact fear AI in finance.
Turning Insight into Action: The correlation between machine learning and job loss is not something that Fleiss.
I asked Fliess “If you’re entering or manipulating data for 80% of your time in a back-office context, might you want to re-think your career?” He states that those are exactly the people he fears for most.
Any company in any industry can ask themselves the following questions:
- What archaic back-office processes do we still rely on today?
- What paperwork and data management processes are still handled slowly and with more manual labor than would be expected?
- For both of the above, where will we move those workers when we inevitably have to modernize / automation those processes in order to stay competitive in the market?
Machine learning, as Alexander sees is, is more about replacing people than augmenting their abilities.
He and I talked openly about how “replacing” is a word that almost no big companies are willing to say. In our past interviews with large companies (such as Accenture, which boasts 300,000 employees), “augmentation” has been the watchword – and for good reason. While I can’t personally speak to the circumstance at Accenture or other large firms, it seems obvious that even if top management was considering job replacement on a grand scale, it wouldn’t be spoken of overtly for fear of terrible backlash from employees, investors, and clients.
By the same token, just as it behooves Accenture to play down “replacement”, it may in someway behoove a smaller startup like Rebellion Research to play up the fears of larger banks, as it may generate more business for leaner more AI-savvy firms like their own – and an astute reader should take strong opinions with a grain of salt. Bias one way or another shouldn’t prevent business leaders from considering the future of the job market seriously.
The following is a condensed version of the full audio interview, which is available in the above links on TechEmergence’s SoundCloud and iTunes stations.
(2.20) Where do you see machine learning currently making the biggest difference in finance?
Alexander Fleiss: There are so many individual parts of finance that are going to be affected, from insurance companies using smart document search to do assessments of people’s risk factors, to doing intelligent background checks….for brokerage and asset management firms. Machine learning is also really going to affect accounting. So often, with inventory accounting, businesses that you deal with…for example, to deal with theft and lost assets, people can use machine learning to see when there are holes in the books within a day or two, without having to wait several weeks.
(6.49) What are some other interesting niche in finance where you’re seeing machine learning creep in?
Alexander Fleiss: One of the coolest things I’ve seen is an Investment Bank in San Francisco, that’s trying to create a machine learning base investment banker. It has a history of past M&A (mergers and acquisitions) and it’s going to focus on industry and it’s going to say something like “Amazon should buy Hewlett Packard….or Garmin should buy some small public tracking satellite company.” The point is, what this bank is trying to build is really a machine learning I-banker, an idea generator for CEOs to make M&A.
(14.02) The bigger hedge funds are starting to pull in additional information and kinds of data when it comes to managing money…all kinds of new streams of data, whether it be Google Earth information that lets us estimate populations or economic activity in different areas where we’re doing business or pulling in social media and headline analysis and aligning that with how we’re going to deal with the candlestick charts in front of us. All of these additional streams of digital data about the physical world and the way that we talk about it are being pulled into the way decisions are being made by people and how they are automated. Are you seeing this trend as well not just with your own work at Rebellion but also in terms of what the big dogs in finance are up to these days?
Alexander Fleiss: Yes, they want to get their hands on as much economic insight as possible. They want the actual raw data, they want to know , for example Michael’s and Co.’s volume in the last week, and the website did X, and what was the market expecting. So from user reviews to any type of volume they can get their hands on.
(20.30) Can you sum up, at least in the finance space…the domains in which you think human effort is most quickly going to evaporate?
Alexander Fleiss: I think we’ll see a lot of brokers losing their jobs, a lot of financial advisors, bankers are going to get hit. I terms of the number of jobs, it’s going to be the retail banks that will fire the most people. They’re going to have fewer people at the window, fewer people in the back office. The big city banks are going to fire tens of thousands of people in operations and accounting; a lot of paper pushers. The banking system of America is a very old school system…and now, with machine learning, over the next decade you’re going to have so many people lose their jobs.
Image credit: Hawk Financial