How AI Is Managing Household Investments and Why It May Not Beat Markets

AI-managed portfolios tend to focus heavily on a small set of popular, large technology stocks, driven largely by media attention and market trends rather than deep analysis. While they sometimes outperform in the short term, they do not deliver consistent risk-adjusted returns, showing little real investing skill beyond following existing market patterns.

Artificial intelligence is no longer confined to chatbots and automation. It is now stepping into the world of personal finance, offering stock tips and even managing portfolios. A new study by the National Bureau of Economic Research, with researchers from Rice University, Michigan State University, and the University of California, Berkeley, explores what happens when everyday investors rely on AI to make investment decisions.

The researchers tested several leading AI systems by asking them to build stock portfolios designed to beat the S&P 500. They tracked these recommendations daily over months, allowing them to see not just what AI chooses, but how it behaves over time.

A Small Basket of Big Tech Bets

One of the clearest findings is that AI prefers simplicity over diversification. Instead of spreading money across many stocks, the systems typically choose a small number of companies, often between 10 and 20.

More importantly, these companies are not random. AI strongly favors large, well-known technology firms. Stocks from sectors like semiconductors and computing dominate the portfolios. In some cases, a single company can take up a large share of the investment.

Over time, this concentration often increases. Rather than diversifying, AI tends to double down on the same companies it already likes, making the portfolio even more focused and risky.

Driven by Trends, Not Hidden Gems

The study shows that AI behaves a lot like a trend follower. It prefers stocks that have recently performed well and avoids those that appear undervalued. In financial terms, this means it leans toward "momentum" and "growth" stocks while staying away from traditional "value" investments.

Even more revealing is how AI decides what to buy. The systems rely heavily on news coverage and online information. Companies that appear frequently in headlines are much more likely to be selected.

This means AI is not necessarily finding overlooked opportunities. Instead, it is focusing on popular, widely discussed companies. In simple terms, it follows the spotlight rather than searching in the shadows.

Does AI Actually Beat the Market?

At first glance, AI portfolios sometimes look impressive. In certain periods, especially when big tech stocks are performing well, these portfolios can outperform the broader market.

But the deeper analysis tells a different story. Once researchers adjust for risk and the types of stocks being chosen, the advantage disappears. The returns are largely explained by the fact that AI invests in high-growth, high-momentum stocks.

In practical terms, AI is not showing true investing skill. It is not consistently picking better stocks than the market. Instead, it is riding existing trends that anyone could follow with a similar strategy.

Active Trading Without Clear Gains

Another important finding is how often AI trades. When given the task of actively managing portfolios, the systems frequently adjust their holdings. This leads to high turnover, meaning stocks are bought and sold often.

While this might seem like smart, responsive investing, it comes with downsides. Frequent trading can increase costs and reduce overall returns. Despite all this activity, the study finds no clear evidence that it leads to better performance.

What This Means for Everyday Investors

The study paints a balanced but cautious picture of AI in investing. On one hand, these systems are fast, informed, and capable of processing huge amounts of data. On the other hand, they show clear weaknesses.

AI tends to build concentrated portfolios, take on more risk, and rely heavily on popular information. It does not consistently deliver better returns once risk is taken into account.

For everyday investors, the message is simple. AI can be a useful tool, but it should not be treated as a guaranteed path to higher profits. Like any strategy, it has limits and risks.

As AI becomes more common in financial decision-making, understanding these patterns will be essential. The technology is powerful, but for now, it behaves less like a master investor and more like a fast follower of market trends.