Outrunning dinosaurs with
artificial intelligence

I am a Millennial, born in the early 90s, which makes me probably one of the last generations to have experienced life before digitalisation took over. Maybe I was too young to be afraid of the millennium bug, but as a kid, using the internet meant monopolising my home’s landline. Digital technologies, once a premium item for few, are today embedded in all aspects of our lives and have brought significant changes and improvements to the way we live and operate. Across the many fields of digital innovation, artificial intelligence (AI) has been the most mysterious and perhaps misunderstood: not just an inspiration for fantasy novels, but a game-changer in information processing, used today mostly in search engines, autonomous vehicles, facial recognition, virtual assistants, and automatic language translation.

Ever since the two letters were first put alongside in the 1950s, AI has seen lovers and haters; those who think robots are going to revolt and destroy mankind and those who think AI is just a cool nickname for using a calculator. The reality is (as always) more boring and way simpler: artificial intelligence is simply a discipline across mathematics, statistics, and information technology that seeks to mimic human cognitive skills within a computer. These skills include learning, processing, perceiving, rationality, and logic.

The investment industry has been trying to take advantage of AI to provide better services, from a customer relation perspective to user experience, or from investment strategies to risk management. In particular, we have been exploring what benefits an AI interface can bring to an investment strategy, especially when integrating digital and analogic, quantitative and qualitative, machine and human brain. A well-trained machine can analyse large datasets and discover new information, change, and adapt to fast-changing dynamics, compensate for human cognitive biases (hindsight, hedging, anchoring, confirmation bias etc.) and for human weaknesses (the struggles with complexity), with immediate application in security selection, asset allocation and research, more generally.

We still have question marks, but the more we learn about it, the more we see the strengths and weaknesses of incorporating artificial intelligence within an investment process. Naturally, there are also some myths and untruths to put right. AI is not always right, especially when the input data is “trash”, or problems are ill-defined; AI is not a black box, or at least no more than a person’s mind; AI is not just a fancy name for a quantitative strategy, as the latter is a human-driven, human-defined set of rules while the former is only a human-supervised autonomous process; and AI will definitely not solve all your problems, as that would be called “magic”.

As an investment manager, we need to keep exploring the extent to which artificial intelligence can enhance our existing investment research capabilities. We need to operate closer to the frontiers of where technology is going. Unless we can tap into that and adopt a more flexible approach to technology, we will lose out as an industry. We risk becoming dinosaurs.