How the investment industry leverages AI

What is AI?

AI is defined as “the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages” (Oxford Languages). You may also have heard the term Machine Learning (ML). ML is a subfield of artificial intelligence but requires human involvement to set up, train, and optimize a computer system. AI and ML have been around for some time, but it is only recently that they have seen such exponential growth, becoming mainstream in the tech world as well as gaining significant traction in several industries which include healthcare, finance, transportation, and retail.

AI and the investment industry

CEOs are now contemplating using AI technology to improve their companies’ efficiency, accuracy, quality, and profitability. As a result, investment professionals deciding on which companies to purchase need to have an understanding of how those companies are leveraging AI and be able to evaluate which industries will benefit the most from these efforts. However, due to the broad and elusive nature of the topic, it is difficult to understand the direct implications for individual companies or investors in those companies, or even how to position a diversified investment portfolio to benefit from these developments. 

In this article, we explore how the investment industry leverages AI and how investment analysts view AI integration at the company level when choosing from several investment opportunities. 

AI and ML are already broadly applied to statistical or econometric modeling techniques within the investment industry. Information processing automation and analysis of large data sets with minimal human intervention have helped analysts and portfolio managers focus their time on more impactful tasks with increasing application over the years. 

More recently, AI applications have been used in some areas of asset management to increase overall efficiency with the intention of improving investment returns. One such area is trading optimization. AI-powered technology can automatically scan the market and provide trading solutions that include the best time, trade size, and trading venues to reduce transaction costs. It can also predict price trends, monitor abnormal transactions, and analyze oversold assets (e.g. when investors overreact to negative news). All this algorithm-generated information can provide real-time data for portfolio managers and traders to enhance investment returns through portfolio positioning. 

AI techniques are also used in portfolio construction and portfolio management by speeding up the risk-return analysis and improving shortcomings of traditional portfolio construction techniques. Risk managers can validate and back-test risk models, generating more accurate forecasts of bankruptcies, credit events, market volatility, and financial crises due to more efficient information extraction, compilation, sorting, and analysis. They can provide outputs based on entire data sets instead of being restricted to conventional sample sizes. 

AI at ATBIM and its sub-advisors

ATB Investment Management (ATBIM) and our parent company, ATB Financial, have always been committed to exploring and embracing new and innovative technologies. We will continue to monitor advancements and continue to explore potential use cases as AI continues to evolve. 

Our in-house portfolio management and investment teams have extensive experience in portfolio construction and asset allocation. Although we embrace technology to assist with our analytical work, we are not using AI to replace critical decision-making functions. We also partner with specialized sub-advisors who bring diversification of thought, style, and experience in a specific area of the capital markets. Our sub-advisors select individual stocks and bonds, and while researching this article, we asked for their perspective on this evolving topic. 

Each sub-advisor confirmed that they analyze the impact of AI on potential investment opportunities but differ on how they leverage it within their own organizations and how they use it to evaluate and purchase securities. All confirmed that AI-related impact is considered when analyzing portfolio holdings or new investment ideas. Technological innovation as a research input is significant when determining a company’s potential as a portfolio holding. 

Our sub-advisors are not expecting to use AI to replace critical tasks currently performed from an investment research or operations perspective. Despite the advancements we see in the headlines, it is not possible to replicate a fundamental bottom-up approach to investing with the AI tools presently available. However, these investment management firms are keenly observing developments in the AI space, with some taking innovative steps to introduce these technological advancements into certain aspects of their workflow.

AI in practice

One of our sub-advisors has successfully replaced some manual work by leveraging AI to analyze company filings using Natural Language Processing, and actively develop and explore ways to complement and enhance their productivity. 

Another sub-advisor set up a research lab that experiments and tests different ways of incorporating computer technologies, including AI, into their investment process without compromising the integrity of their proven process. As part of these tests, they developed an automated ML tool to analyze historical financial information and pre-populate a discounted cash flow model. 

The research lab has developed another tool that conducts credit evaluation of securities using synthetic credit ratings where public ratings are unavailable. As the results from these efforts are compared and validated against traditional analysis methods, they will continue to test their reliability before incorporating them into their decision-making models. 

Risks & Potential Concerns

When considering using AI in investment decision-making, being mindful of potential risks is essential. “Black-box” and complex AI models are difficult to scrutinize, and caution is required when taking their output for face value. The concept “garbage in, garbage out” applies here— these models are only as good as the source data used, making them sensitive to poor-quality or insufficient inputs. There is also potential for so-called “hallucinations,” when a model can randomly create data or facts, and “degradation,” where a model performs less optimally and provides poorer quality output over time. Models need to be rigorously validated and tested before being relied upon. While there are still worries that AI could replace human talent in the long run, another perspective is that AI applications will allow human talent to refocus efforts more productively and efficiently, eliminating the need for human intervention in repetitive and mundane tasks. 

On a cautionary note, as with all significant societal disruptions, adverse impacts cannot always be anticipated, understood, or priced into valuation models. For example, when the internet rose in popularity and became mainstream in the late 1990s, many tech stocks became overvalued. There were also critical unanticipated negative consequences for vulnerable users, and the risk surrounding the hype of promising internet companies burned many shareholders during the dot-com crash. 

The current excitement surrounding AI provides perfect breeding grounds for potential fraud, illegal usage, and AI models producing unethical output. Investors are also not immune to repeating history by inflating the stock price of any company they perceive as the beneficiary of AI advancements. Information security and privacy concerns are another major issue associated with using AI-powered solutions, including open models (such as ChatGPT). Skepticism is not uncommon, even from technology leaders. In his seven-page letter, “The Age of AI has begun,” Bill Gates pointed out that this revolutionary technology presents other dangers, such as AI tools ignoring directives and pursuing their own goals.

What does the future hold?

AI has already had a profound impact on many aspects of our lives. We may not recognize the world as we know it in the next five to ten years as AI becomes more widespread, realizing its potential, and more industries and companies leverage it to improve workflows. AI is a great tool aimed at improving productivity, but caution needs to be exercised with when and how to use such tools, including when it comes to money management. While the AI hype is very alluring, more time is needed to fully understand the implications of the newer aspects of this technology.