As students everywhere head back to school, the possibilities presented by artificial intelligence are driving fear among teachers and professors, yet excitement in the stock market and in a wide variety of commercial and research efforts.
For institutional asset allocators, a similar mix of anticipation and skepticism exists regarding what AI, and the large language models that inform it, can bring to investment processes and organizations.
The role of AI and LLMs in the future of institutional investing is going to be significant, but for the foreseeable future, it is seen mostly as a tool to aid human investors, not replace them, according to CIO NextGen honorees’ views on AI’s involvement in investment strategy.
Adrian Ohmer, director of investments at the Kresge Foundation, likened AI to another powerful technology that changed the way investors have operated over recent decades.
“I see a very real path that AI and other large language models can be the next Bloomberg of our industry,” Ohmer wrote in his response to CIO. “That said, the question becomes whether currently existing service providers and tools harness this power to make their tools even better for users or if there’s some truly disruptive new entry into our market.”
Joshua Adler, senior manager of hedge funds and portable alpha at Raytheon Technologies, said the impact of the technologies will be “profound,” but that today, AI is a “convenient tool. However, it will change the way we invest going forward from how data is synthesized and analyzed to creating efficiencies within the due diligence process and developing a sound investment thesis.” Still, he added the caveat that “it will take time before we trust AI with decisions regarding strategy and market insight.”
Han Yik, a senior adviser to the executive director and CIO at the New York State Teachers’ Retirement System, acknowledged that the technologies have a place in investor organizations.
“AI has the potential to be an extremely powerful tool and will be able to analyze rapidly changing economic conditions … but ultimately, I see us still utilizing it as a tool, with human oversight, much as we have gotten used to employing other disruptive technologies, such as computers and the internet, in our businesses,” he wrote.
Charlotte Zhang, investment director at the Inatai Foundation, cited the utility of using algorithms to reduce some workloads for investors, who can then focus on more valuable decisions.
“When evaluating a deal, human cognition exhibits limitations in processing, continuously tracking and assigning weights to the volume of information across comparables,” Zhang wrote. “Therefore, it would seem logical to leverage algorithms to automate collecting, analyzing and presenting data in a standardized, digestible way. This frees up VCs to utilize more of their bandwidth on … [efforts] that enable higher-conviction decisionmaking.”
Because AI and other technologies are good for the number crunching and analysis that many investors and allocators do early in their careers, the honorees had different thoughts on what the tools may change within organizations and what they do not want to see lost.
“When the right questions are asked and algorithms are correctly configured, AI holds the potential to reduce the time required to develop expertise in a specific subject and shorten the learning cycle,” wrote Brandon Tasco, manager of real assets at the UAW Retiree Medical Benefits Trust. “Our business, like many others, relies on obtaining direct experience and knowledge, which used to only come from an apprentice approach and getting passed-down expertise. … By harnessing AI and large language models, we can accelerate the learning cycle, resulting in better investment decisionmaking.”
John W. Pearce, portfolio manager at the Illinois Municipal Retirement Fund, pointed to the importance of the analytical work done by an early-career investor as a key to later success.
Because generative AI has the potential to supplant young professionals in doing basic modeling and coding for investors, Pearce wrote, “The opportunity to frugally expedite this work is offset by the risk that the next generation of investors will miss out on critical repetitions in analysis that provide the experience required to develop deep understanding. Aspiring musicians practice scales and études because mastering these building blocks allows them to be able to improvise and create new sounds entirely. … I fear that being overly eager to embrace generative AI for the foundational exercises of investing risks trading off near-term gains for long-term talent development problems.”
Other NextGen honorees pointed to the responsibility allocators have to the pensioners and other beneficiaries of the funds they manage and that AI would not mitigate those obligations.
“It will be crucial to take a measured approach in layering in these applications—as fiduciaries and stewards of long-term capital, allocators will have to understand the rationale of any AI system’s decision-making framework in order to incorporate its output in a potential investment-related decision,” wrote Thomas Kim, an investment officer at the San Bernardino County Employees’ Retirement Association.
Pearce cautioned against expecting too much from the new tools. “While these methods provide an opportunity to uncover gold, there is the ever-present danger of data mining,” he wrote. “Recent advances in technology have upgraded the pickaxes of yore for dynamite, exacerbating the risk of finding mere pyrite,” otherwise known as Fool’s Gold.
Steven Kaell, managing director at Memorial Sloan Kettering Cancer Center, acknowledged the important efficiencies that come with AI and other technologies, but also cautioned that there are limits. “While leveraging these immediate benefits, we need to be discerning consumers of the output, which may be inaccurate or lack nuance,” he wrote.
Carmen Lugo, a portfolio manager at the Fondo de Ahorro de Panamá, put it succinctly, writing that “human … expertise and judgment will remain essential and cannot be replaced.”