A leading figure in quantitative investing has issued a warning against fully delegating investment decisions to artificial intelligence, suggesting that certain segments of the hedge fund industry may have adopted the technology too aggressively. Martin Lueck, one of the three founders of the pioneering quant firm AHL and co-founder and president of hedge fund Aspect Capital, stated that the ability to accurately understand why his firm's computer models recommend specific trades is of critical importance. "My starting point is that I will not put my name or my firm's reputation behind something where I have no clear idea why it holds certain positions," said Lueck, whose firm manages $9 billion in assets. "I need to have a hypothesis I can rely on. If I were investing my own money, I would want to know what it is doing," he added. Lueck's comments come as hedge funds and proprietary trading firms deepen their use of AI and machine learning technologies. Cliff Asness, billionaire founder of US quant giant AQR, stated last year that his firm was "ceding more decision-making to the machines." Asness acknowledged that his fund is capturing patterns that its quantitative researchers sometimes cannot explain and is using machine learning to determine how much capital to allocate to trades. "After a very tough period, this has been very good for us, which makes it easier," Asness said at the time. "It will likely be harder to explain to investors during bad times, but we believe it is clearly worth it." Lueck said he has heard Asness's perspective and finds him "very persuasive." However, he added that one key reason he left Man Group—which had acquired AHL—back in 1995 was precisely to offer investors greater insight into the factors driving trading models. "Back then it was a black box," he said. "There was zero transparency into what these models were doing; the entire quant world was opaque." Traditionally, quantitative investing uses deep research—often drawing from academic papers—to identify and explain market patterns, which are then traded by computer algorithms. Such programs typically include rules governing capital allocation and leverage adjustments based on market risk. Lueck, who co-founded AHL with Michael Adam and David Harding, was an early pioneer in trend-following strategies—a branch of quantitative investing that aims to profit from investor herd behavior in buying or selling certain assets. However, with the rise of greater computing power and increasingly sophisticated AI and machine learning, hedge funds are tempted to delegate more responsibility to machines. Lueck added that there are significant opportunities in using large language models to assist quantitative researchers with tasks such as organizing data, running tests, and preparing presentations for colleagues. But he also cautioned: "I still want researchers to think about 'what exactly am I researching,' rather than 'here is some data, find me a relationship.'"
Comments