Opinion: Finance Looks Like a Sector Built for the AI ​​Revolution

Gus Carlson is an American columnist for The Globe and Mail.

For the past few years, some of the best jobs in tech haven’t been in tech at all. They’re in financial services.

Banks, investment houses, hedge funds and private equity firms are recruiting tech talent like crazy, eager to get a jump on fast-growing technologies like artificial intelligence, using lucrative compensation packages and Wall Street lifestyle perks to lure data scientists, software engineers, business systems analysts and others from the tech sector, and to poach big brains from top tech schools like MIT, CalTech and the Ivys.

The Street got a taste of the benefits this week. Bridgewater Associates, a large Connecticut-based hedge fund, launched a $2 billion fund that uses machine learning to make decisions, using models developed in partnership with AI leaders including OpenAI, Anthropic and Perplexity.

Although Bridgewater claims that the fund will be overseen by humans, make no mistake, this is a watershed moment. Technology, which has slowly but surely displaced human talent in the financial industry, has broken through. Think Terminator 2 meets Wolf from Wall StreetThe machines win.

The nagging question is: Who cares, as long as the fund makes money? Well, for those employed in a wide range of roles in the investment industry, that’s a wake-up call.

McKinsey’s 2022 prediction that 30% of human hours in financial services would be eliminated by 2030 now seems too modest. In fact, it could be much higher if funds like Bridgewater deliver great results and create more like it. No wonder research shows that more than half of all finance workers believe technology puts their jobs at risk.

Bridgewater, which has more than $100 billion in assets under management, could be a model for the future. Executives have already said the push into machine learning will likely mean a greater focus on hiring people like data scientists, further shifting the balance of the workforce toward technology expertise rather than financial expertise.

Like many financial services firms, Bridgewater has been driving its technology transformation in-house for nearly a decade, with major hires from the tech world—a senior fellow at IBM Watson—and from academia—a statistics professor at Yale.

The firm also created an AI skunkworks to develop the proprietary systems that will underpin the new fund. To make sure the technology works, it was tested in an existing fund over the past year. While past performance is no guide to future success, it’s worth noting that a fund with embedded machine learning delivered double-digit returns after years of mediocre performance.

While Bridgewater’s fund certainly has a wow factor, that shouldn’t come as a surprise. It’s a natural evolution of the business, especially given the investments in technology made by the major players. And honestly, there’s probably no better sector than finance for AI to show its capabilities.

Machine learning and AI can do amazing things, and do them quickly, in the money business. They can identify and exploit trading patterns in the blink of an eye. They can see cause-and-effect relationships between markets, and they can scan, assemble, and collate company reports, documents, and announcements, media headlines, and analyst reports from around the world in a nanosecond.

Of course, there are some things AI can’t do and never will be able to do. Namely, it can’t pick up on the whispers that often mean the difference between getting in at the right time and, just as importantly, getting out.

AI doesn’t sense the nervousness of a CFO or CEO on a quarterly earnings conference call or an investor day presentation trying to explain falling earnings per share or unexpected spending on initiatives. It doesn’t have the inside scoop on the unofficial, behind-the-scenes, deep dives into a company’s strategy or the competence and focus of its leadership. These are the nuances that don’t show up in bits and bytes, but often give investment professionals and their clients a competitive advantage.

Even Bridgewater co-chief investment officer Greg Jensen, who will oversee the new fund, told Bloomberg that big language models “have a hallucination problem. They don’t know what greed is, what fear is, what the likely cause-effect relationships are.”

But the truth is that if this new fund generates healthy returns, conventional qualitative and contextual insights likely won’t matter—or at least will become subordinate in terms of value. Just as people don’t care whether their favorite Netflix shows were created by humans or AI as long as they’re entertaining, investors won’t care how the fund performs as long as it does.