Where Will Investment Edge Be Found in the Age of AI?

AI is rapidly changing how investing is done.

Investing involves finding information, analysing what it means, making a judgement and ultimately acting on it.

Until now, AI has mostly helped with the first two. The investment decision itself has generally remained with the investor.

Questions such as how much of a company’s earnings are exposed to one country, how sensitive a business is to interest rates, or what new technologies could disrupt a particular industry previously required searching through multiple reports and filings.

AI can often get much of the way to an answer in seconds.

But more information and more sophisticated analysis do not automatically lead to better judgement.

One recent study found that AI significantly increased both the breadth and depth of stock analysts’ research: they used 40% more information sources, covered 34% more topics and employed 25% more advanced analytical methods.

Yet their forecast errors increased by 59%.

In other words, better research did not necessarily lead to better investment conclusions.

A similar picture emerges from recent research that asked AI models to pick stocks. After adjusting for risk, the models did not generate statistically significant abnormal returns.

This suggests that, at least for now, AI is commoditising analysis faster than judgement.

But there is no reason to assume AI will remain weak at judgement forever.

AI may eventually become better at asking the right questions, weighing evidence, and learning from past mistakes than humans are. If that happens, judgement itself may become less differentiated too.

But investing has never been only an analytical problem. It is also a psychological one.

Knowing that an unpopular investment is attractive is different from being willing to own it while everyone else disagrees.

If AI makes information, analysis and eventually even judgement increasingly commoditised, the durable edge may lie less in knowing more than everyone else, and more in being willing to act differently.

This is one reason why I believe in investing in the unpopular. Markets will become better at processing information as AI improves, but I don't think human behaviour will change nearly as quickly.

There will still be investments people don't want to own, narratives that become too pessimistic, and periods when following the crowd feels much more comfortable than going against it.

Being willing to invest in those situations will remain a source of opportunity, even in a world where everyone has access to powerful AI.