Although AI (Artificial Intelligence) has been around since the 1950s, the recent explosion in the use of generative AI chatbots like ChatGPT and Bard has put it literally at the fingertips of millions of ordinary people in their everyday lives.
Already ubiquitous in areas such as voice and face recognition on our mobile phones, the pace of technological change means that industries and governments are scrambling to come to terms with what AI might mean and how to exploit its potential while managing the risks.
One person with “real hands-on experience” of AI is Dr Tom Humberstone, a Research Intern on the Courtiers Investment Team. Tom says he’s used AI both in academia when completing his PhD at Oxford and while working with Courtiers CIO Gary Reynolds at Rams Rugby Club.
With its ability to crunch huge amounts of data and reveal patterns and links that would otherwise remain hidden, Tom says AI was able to help identify that the underlying reason one Rams player wasn’t performing to his potential was because of an undisclosed injury, which was then able to be treated. “There’s no way humans could have got to that, there’s way too much data,” says Tom.
With its focus on data and the importance of being able to draw useful insights from that data, investment management is on the face of it a ‘natural home’ for AI. Given Tom’s experience of using AI algorithms to mine and create data and make predictions, I was keen to get Tom’s views on the role AI might play in the financial services sector and within Courtiers.
In the context of investment management, Tom was quickly able to roll off a number of areas in which AI might play a useful role. One of these could be to find companies, which show up very badly on standard screening measures, but where AI could pinpoint the reason as “simply because of one outlying negative bit of data that might actually be irrelevant or fixable.” It could also be used to find previously hidden undervalued assets.
When it comes to AI at Courtiers, Tom says he’s working to improve a tool used called ISRRA (Investment Strategy Risk Return Analysis) that has elements of AI within it. By tweaking the algorithm that calculates expected returns, he says the work when completed will help Courtiers Advisers produce the optimum cash drawdown strategy for clients under different market conditions. ISRRA aside, he says the company is at “the exploratory stage”, conscious that new technologies are often initially overhyped before settling down. That said, within the Investment Team, there are “very active conversations going on about the long-term use of AI, its value and the risks of using these tools.”
Tom says that AI is especially suited to dealing with what’s called “high dimensional data”. The human mind “doesn’t have the intuition to work with 30 or 40 dimensions at the same time comparing and evaluating them,” he explains. Another potential use of AI could be to extract self-declared risk profiles from company filings. This would allow investment managers “to quantify the risk in different sectors, and to understand the relationship between risk and excess returns and the nature of that risk itself.”
Managing the risks
Tom argues that the ability of AI to give “really specific insights” into complex subjects should over time as AI spreads across the sector, lead to a general rise in performance across the industry. Although clearly excited about the potential uses of AI in the finance sector, Tom warns against leaning too heavily on algorithms and statistics.
He’s clear that the interpretation of what AI produces and decision making must remain the prerogative of human beings. Tom highlights the risk of what’s called ‘blackboxing’, where what goes into an AI algorithm is effectively invisible and where decisions or conclusions are arrived at without providing any explanations. “There could be historical biases tied up in it, there could be huge blow-up risks, which AI doesn’t understand, but which are obvious to a human being.”
Tom accepts that as AI tools become more readily available across the finance sector, it will be harder for companies to retain a competitive edge. He argues that in the long run, it’s going to be those who think most sensibly about AI and how “these amazingly powerful tools” interact with human decision makers under risk and uncertainty, that will have the best chance of delivering superior benefits to clients.
With an explosion in the use of AI expected in the sector, Tom says the role of regulation to protect consumers will be crucial. He says he’s hopeful that the regulators will build on earlier work, which included a helpful report published by the House of Lords in December 2017.
While only time will tell, Tom says he believes that those companies who take “a cautious thoughtful approach” will find themselves in poll position. The future of AI will be about “skilled practitioners leveraging or ‘co-piloting’ these tools to get those really powerful insights rather than using AI to replace everybody.”
While purporting himself to be “incredibly excited” at the prospect of getting access “to some of the most powerful tools ever invented”, Tom refuses to get carried away. Yes, there’ll be things that will change, but as he’s keen to point out, Courtiers has always been about personal relationships, particularly between advisers and clients. Even with the inevitable march of AI, by focusing on using the power of AI to enhance those relationships “I don’t think that’s going to change anytime soon.”