Artificial intelligence may one day upend whole industries, but among the hedge fund elite, it’s not yet a threat, investing legend and Baupost Group CEO Seth Klarman said.
In an episode of Columbia Business School’s Value Investing with Legends podcast this month, the hedge fund titan said he and his colleagues mostly use the nascent technology as a time-saver, assigning it tasks that lower-level staffers may have done in the past.
“We have ourselves started using it as essentially a capable assistant, a summer intern,” Klarman said. “Not somebody who knows which stocks to buy, but a way to tabulate data quicker.”
Klarman cited a couple of specific examples of how his colleagues have used it.
One way was to analyze corporate filings.
“He asked AI to take a look at 10 years of annual reports for a company and compare what changed in the annual report from year to year — the way they communicate, which may reveal something about the change in the business, or what the lawyer was worried about, or whatever it might be,” Klarman said.
Another was to identify company logos that one of his employees had never seen that would help them know an industry landscape better.
“He would have run that to an intern, but AI did it in like five minutes and saved three days’ worth of work,” Klarman said.
Klarman also said he’s been using AI himself, though he hasn’t been particularly impressed. Before he was set to speak at an event with a well-known business executive, he asked a chatbot to generate some questions to ask the person.
“What came back was useless,” he said. “I like to ask things that they haven’t thought about before.”
Baupost isn’t the only firm using AI to increase efficiency. Morgan Stanley and Bank of America have started to train staff to use the technology to boost productivity. Consulting firm ThoughtLinks found in a recent study that by 2030, 44% of a bank’s tasks, including 32% of sales and trading work, could be “redefined” by AI.
Despite using the technology, Klarman said he’s concerned that AI could hurt people’s ability to be creative. A recent MIT study backed this up, finding that using AI reduced memory and brain activity.
“I think that if we use AI the wrong way, we’ll solve the problem without having applied our brains,” Klarman said. “It would be like reading the end of a novel and then not needing to really know as well how did the whole thing come about? So I like doing it in order. I think the right order is, ‘Here’s what I think,’ and then I can improve my thinking.”