Wall-E Street

The big story this year not only in equity markets but also in society has been Artificial Intelligence’s (AI) dramatic entrance onto the stage. Since then, a deluge of headline-grabbing commentaries have flooded in, speculating on various implications of this new technology. The mania left me with some concerns: Will AI pave the way to apocalyptic destruction? What does it truly mean to be sentient? But most importantly, will I lose my job? While predicting the end of life on earth is perhaps beyond my pay grade, I have written some thoughts down on AI’s role within the investment management industry and whether I should start looking for a new job.

AI, broadly defined as the ability of a machine to problem solve in ways demonstrated by humans, is nothing new to the world. We have consistently been caught off guard estimating AI’s abilities, beginning with the defeat of world-renowned Chess grandmaster Gary Kasparov at the hands of a supercomputer, Deep Blue. Since then, AI has worked its way into everyday life in ways such as providing uncomfortably relevant advertisement recommendations based on past behaviour, employing voice assistants like Siri and Alexa, and now ChatGPT.

AI is also no novelty to asset managers who have been conducting quantitative analysis through complex computer programmes for many years. So, what has changed? Well, in recent years there have been meteoric advancements in computing power, alongside an explosion in both the volume and pace of data output. Consequently, AI models can now analyse larger datasets in a shorter amount of time. By sifting through swathes of data, computers identify patterns and generate enhanced predictions far quicker than mere mortal analysts. The newer models are also able to autonomously adapt to changing market conditions, removing the need for tedious re-calibrations.

Another development has been Natural Language Processing (NLP) – a subset of AI which allows computers to extract meaning from natural language and speech. Such techniques permit access to a broader spectrum of data sets, known as alternative (or unstructured) data, opening possibilities for unique and differentiated insights. This encompasses sources such as tweets, earnings calls, and credit card transactions, differing from traditional data such as financial statements or official economic figures. The use case is evident: greater diversity and volume of data leads to more-informed investment decisions.

However, a crucial aspect of investment management involves cultivating personable relationships with clients, built on trust, but most importantly accountability. The complex nature of AI models makes it tricky for the few investment managers not well-read in data science and machine learning to explain the inferences drawn from them. The inability to explain performance attributions and investment decisions to clients, who entrust large proportions of capital, undermines the fiduciary duty and trust nurtured over many years.

Adopting such complex strategies also entails substantial costs and a shift away from established processes designed to leverage the know-how of the investment team gained from years of experience. One option, however, is to partner with a third-party entity that already possesses the requisite expertise and infrastructure, thereby gaining access to the rewards of such technological capabilities without the additional financial and time-related burden. Beyond the allure of convenient sounding names, this was the rationale behind MGIM’s recent partnership with AI-focused fintech company, MDOTM. MDOTM conducts macro, fundamental and price-based data analysis to provide asset allocation insights and suggestions through their proprietary AI engine ALICE (*ominous villain music*) – the newest member of the MGIM team. ALICE will also be able to tailor her outputs to work within client’s risk appetite and given asset class weighting constraints. What ALICE lacks in her contributions to office banter, she more than compensates for in her efficient and unbiased analytical skills.

Nevertheless, this primarily serves as a supporting function intended to complement our existing strategic/tactical asset allocation and portfolio construction processes rather than replacing it. Ultimately, the final decision will remain in the hands of human expertise which clients pay fees for and value, something which cannot be replaced by algorithms. In conclusion, for now, I do not foresee ALICE as the evil antagonist poised to forcefully remove me from my desk, rather an amicable colleague, bringing an alternative perspective. So, ALICE, while you may not be on the invite list to my wedding anytime soon, I look forward to our time working together.