Camille Van Vyve
02 Jul 2025
Can AI beat the market? Separating hype from reality
Artificial intelligence is reshaping every industry, and the world of investing is no exception. Tools like ProPicks AI now claim to "beat the S&P 500" or "outperform the Dow Jones" using AI's analytical power. These bold promises raise a fundamental question: can AI really outperform the market consistently over time? Let’s take a closer, more nuanced look.
Algorithms and AI: impressive capabilities
AI has clear strengths when it comes to financial analysis. Its ability to process massive amounts of market data—financial reports, economic news, sector trends, macroeconomic indicators—in seconds far exceeds human capacity. “Even before AI came on the scene, we’d already come a long way in terms of speed and complexity of data processing,” notes Roland Gillet, professor of financial economics at the University of Paris 1 Panthéon-Sorbonne and Solvay (ULB), and advisor to both public and private institutions. “What sets AI apart from traditional algorithms is its adaptability—its ability to improve on its own. But even with an intelligent model, the future will always be, in part, unpredictable.”
Hidden correlations—or random noise?
Gillet illustrates this point with an anecdote from a 1989 research seminar at Harvard: “Two well-known researchers were trying to prove that stock prices don’t follow a random walk—that historical data contains correlations that could predict future prices. They aimed to identify ‘pockets of inefficiency’ in the New York Stock Exchange. During the session, an attendee handed them a disk with stock price data and asked them to identify any correlations they could find. They did. Then the attendee revealed that the data was entirely random, generated by a simulation. The researchers had uncovered patterns that, in theory, shouldn't have existed—except by pure chance.”
The human (and emotional) factor
Financial markets aren't purely rational systems. They’re heavily influenced by human behavior—herd instincts, fear, euphoria—which can create irrational swings that are hard to model. AI can analyze past data and simulate scenarios, but human reactions to new events remain wildly unpredictable. “When Trump says one thing and then the opposite, markets need time to adjust,” explains Gillet. “They’re likely inefficient for several minutes or even days. AI can’t foresee where his behavior will lead next.”
The problem of extreme events
So-called “black swans”—a term coined by Nassim Nicholas Taleb—are another hurdle for AI. These are rare, unpredictable events that have major consequences for markets: large-scale natural disasters, major tech outages, the sudden death of key CEOs, or unexpected geopolitical shifts. They’re, by definition, impossible to predict using historical data.
Faster data integration—an edge?
Could AI gain an advantage by processing available information faster than anyone else and identifying opportunities before the rest of the market catches up? “It’s possible that AI might process information even more quickly than other forms of intelligence, at least over very short timeframes,” Gillet concedes. “But even if that’s true, I believe the impact will be fairly marginal.” According to market efficiency theory, any successful AI strategy would soon be copied by others. As more investors adopt similar models, the edge would disappear.
The future of AI in investing: human-machine collaboration
1. Smarter decision-making
Like previous algorithmic models, AI can quickly analyze large volumes of data and highlight potential investment opportunities—opportunities that human analysts can then assess using their own judgment and expertise.
2. Better risk modeling
Machine learning algorithms continuously improve risk models by integrating new data and adjusting parameters dynamically.
3. Democratizing financial analysis
AI-powered tools give individual investors access to sophisticated analysis that was once limited to large financial institutions, helping to make markets more accessible to everyone.
AI at Easyvest: automation that serves our clients
1. Smart algorithms designed at Harvard
Since day one, we’ve based our portfolio allocation and rebalancing on unique algorithms developed at Harvard. These mathematical models help us optimize diversification and manage risk—but they don’t claim to predict the market.
2. Automation of low-value tasks
We automate as many administrative and back-office tasks as possible, allowing our wealth managers to focus on what truly matters: understanding our clients' goals and delivering personalized advice.
3. AI-powered chatbot
Our app now includes an AI chatbot trained to answer nearly all your questions about investing and Easyvest. This virtual assistant complements—never replaces—our human experts.
Conclusion: AI as a partner, not a substitute
AI is undoubtedly transforming investing. But claims of consistently beating the market through AI should be met with healthy skepticism. Markets are shaped by human behavior and unpredictable events—elements even the smartest algorithms can’t reliably foresee. At Easyvest, we believe in a balanced approach: leveraging the power of AI and automation to improve operational efficiency and analytical accuracy, while keeping human expertise at the heart of the client relationship. This synergy between technology and human judgment is, in our view, the best way to navigate the complexity of today’s financial markets.