The evolution of AI and machine learning have re-imagined our perspective with respect to most spheres of life. There are differing views on AI with some viewing it as a technology that has the potential to change the way we live and work while others see it as a threat to our privacy and security. And others still who are unsure of what to make of it.
Despite the mixed sentiments, one thing stands clear - data is the life blood of AI and machine learning.
A fundamental question that many ask themselves is whether the output from AI can be trusted. Most recently conversational AI like Chat GPT have been making the news. However conversational AI does run this risk of being bias based on the training datasets that it is fed. In simple terms, if the data sets included a significant number of article that were say from a far-right wing lens, the the AI output too would seem right wing.
So is the use of machine learning with respect to investment strategies any different from conversational AI? The answer to that is yes, absolutely. The most stark difference is that there is no ambiguity in the data sets. With respect to portfolio construction, the data is fed directly from the stock exchanges and unlike conversational AI, these algorithms are objective and based on pure math.
The more pressing question is with regard to the debate on man vs machine. With the increase in both the available data and analytical and computing horsepower, does it mean that the traditional role of a wealth manager is completely obsolete? To understand the magnitude of the role machine learning has to play in portfolio construction let’s first look at the data available to a wealth manager. If you consider both the direct feed from the exchanges (as well as derived metrics) that is over 1 million datapoints. To put that in perspective, if each data point represented an inch-long piece of string, lined up they would measure 25.4 km - that’s roughly the length of 56 football fields laid end-to-end. Now imagine this wealth manager trying to draw connections between them and make dynamic changes to it in line with needs of multiple of clients. It’s just not humanly possible.
Modern computing power and machine learning is able to identify deep connections within the financial universe building a neural network of sorts, invisible to the human eye. And with more data and more time, this understanding only gets deeper and more accurate.
There is truth in data and with the right approach, the potential for artificial intelligence in wealth management is limitless.
That being said, the stock market does not exist in a bubble and there will always be room for nuanced human intuition, especially so for outlier situations such as corporate governance issues and black swan events.
Moving forward, the best positioned financial advisors would be those that harness the power of data and augment their conventional, fundamental based approach with machine learning.
