Ten years ago, if Breanndán Ó Nualláin told people that he is a professor in AI, they would not really have an idea what that was all about. This has radically changed. Nowadays every newspaper will carry some article or mention of the topic. All this attention notwithstanding, finance professionals still struggle to grasp the enormous potential AI offers their field, says Ó Nualláin. He teaches the course AI for Finance Professionals at the Amsterdam Institute of Finance.
One of the most straightforward applications of AI is in financial reporting. Large amounts of text on the internet have already been written by tools like ChatGPT. Starting from the financial data, AI is able to write summaries, Ó Nualláin explains. “Say you want to write a one pager, a good place to start might be to ask ChatGPT to give you an outline. Then you have something to get going, you don’t start with a blank page.”
AI as a junior
AI could actually write whole parts of the reporting, he says, adding an important caveat. “We see people saying they want it to write a report. This can be done; however, you can’t just take the output of the model, put a cover on top, and declare it finished. You should be really critical about the result in the same way as you might be critical of a junior. You need to examine the output carefully and use it as a first version.”
When you don’t use data, you’re sitting on an asset
Ó Nualláin strongly encourages finance professionals to look beyond these basic applications. A good starting point is to see data as a resource. “Ask yourself what data is available in your organization that is not being used. Data is an asset. If used in the right way, it can be very advantageous. If a company owned a large building and left it empty, people would ask questions why you’re not using it. In the same way, when you don’t use data, you’re sitting on an asset.”
When it comes to data analysis, humans have limitations that AI doesn’t have, Ó Nualláin explains. “Typically this is data that is highly dimensional, there are many, many columns. We as humans can’t think in more than three dimensions, but there are many techniques to slice and dice data to give you actionable insights.”
Something that’s done in market analysis is segmentation, he elaborates. “We look at different groups of society, like the sporty young men or the soccer moms or the high net worth professionals, but then you’re painting with a very broad brush. Using machine learning techniques, we have the ability to zoom into much smaller clusters of people.”
AI can endlessly slice and dice data
He gives an example from the insurance sector. “So, someone wants car insurance. It’s a man between 25 and 30, living in Amsterdam. Traditionally, actuaries would draw up tables based on that information. However, the risk would be different if that person had a job in night life and worked late nights than if that person were a university student. The risks for these two classes of people are different, but an actuary would take them all with the same brush. Machine learning can give us these insights.”
Fear of losing control
The many potential applications notwithstanding, many finance professionals are quite reluctant to educate themselves about these techniques, let alone apply them in their work, Ó Nualláin sees. “A majority really doesn’t have more than a passing knowledge of how this works. They have vague, sometimes strange ideas, that often lead to outsized expectations or exaggerated fears or apprehension.”
He sees a fear of losing control when data and systems are moved to the big organizations that are at the forefront of the AI-revolution. “It’s a feeling of being lost, of not knowing what’s going on when we put our systems and data in the hands of people who are doing all this.” However, this is just another argument in favor of understanding what AI is, how it works, and what it can and cannot do. “When you’re using this technology, it’s critical that you have a clear picture of what is good, and what is happening behind the scenes.”
Edge
Ultimately finance professionals would be well served if they gained the basics of AI, Ó Nualláin concludes. “I think it would be very valuable for them to educate themselves a little bit more about AI so that they can at least talk to the professionals who are doing AI and machine learning, and be able to do it in a well-informed way. It certainly gives you an edge when you’re able to have these conversations.”