10 May 2019
The financial services industry can’t ignore the impact of AI on investments any longer. It’s here to stay, bringing with it a whole new way of doing business, as evidenced by the World Economic Forum 2019, where the topic of AI dominated discussions.
Some issues to consider are how to integrate AI into the investment process, whether to use full or partial AI, and the future role AI will play in the process of conducting a due diligence on asset managers. So we need to understand its impact on the industry as a whole.
AI is broadly defined as the use of machines to perform tasks that would normally be done by humans. Ultimately, AI seeks to create systems that can function intelligently and independently of humans.
Many people use AI daily, often without even realising it. Examples include Google searches and using a smartphone’s GPS.
AI works in two ways:
Most asset managers are currently operating in the narrow AI space, but increasingly moving towards deep AI, and to a future where machines may even have self-awareness.
In a world of big data, where we create up to 2,5 quintillion bytes of data every day (that’s 25 and 20 zeros), AI is a non-negotiable. Humans learn in 2 or 3D, whereas machines are able to learn in dimensions that humans cannot begin to comprehend, 100 or even 1 000D.
Having mastered pattern recognition (seeing patterns that humans can never see), the next steps in ML are categorising information and then making predictions that humans can never make.
AI reduces a number of key risks in the investment process. It eliminates human bias and emotion, mitigates key man risk, fills the gap in under-resourced investment teams, and smooths out inconsistencies in the investment process – all while dealing with huge volumes of existing data and the rapid growth of new data.
AI processes many different categories of big data. These include unstructured information in the form of web searches, sentiment and social media commentary, and structured information, comprising macroeconomic data, asset prices and financial statements.
Over time, structured and particularly unstructured data grows exponentially, while the level of human ability to process remains largely constant. That’s why we need AI and ML to be able to either make or help us to make investment predictions based on large amounts of data.
The Glacier AI Flexible Fund of Funds is one of the very few funds worldwide that employs a full AI strategy using learning algorithms to run the entire investment process – from idea generation right through to making the final investment decision.
Glacier’s AI engine, also called PIE (Predictive Investment Engine) focuses solely on structured data to make its investment predictions, which we believe currently ensures the best performance from full AI engines. However, when using both structured and unstructured data, partial AI (combining both humans and machines) will more than likely do better than full AI – but that is only until AI technology in investment management is limited to using narrow AI Things may change if deep AI is used more broadly in the future.
Many investors construct portfolios taking investment style into account. However it’s not easy to equate full AI with any particular investment style – such as value, growth, momentum or quality – as full AI is adaptive to different market environments. A core (neutral) investment style would potentially be the closest style comparable to full AI.
Full AI offers these diversification benefits in an investment portfolio:
Warren Buffett once said, ‘You don’t need a lot of brains to be in this business. What you do need is emotional stability. You have to be able to think independently.’
That could prove the best unintentional endorsement yet for the inclusion of AI in an investment portfolio.