Has AI gotten any better at stock picking?

Artificial intelligence has been touted as a transformative force in finance, and the question of whether it can pick stocks better than humans has drawn significant attention. The recent data, however, points to a sobering conclusion: beating the market remains remarkably difficult, even with sophisticated AI models at your disposal.

The core challenge lies in market efficiency. Prices already reflect widely available information, including much of what these models are trained on. An AI system can identify patterns in historical data, but future market movements are driven by unpredictable events, shifting macroeconomic conditions, and the collective behavior of millions of participants. Simply processing more data faster does not guarantee an edge; it often just means the model is learning from noise rather than signal.

For investors, the lesson is not to dismiss AI but to calibrate expectations. The value of these tools may lie more in improving research efficiency, risk management, and portfolio construction than in generating outsized returns through stock selection. What matters is how the technology is integrated into a disciplined process, rather than any promise of guaranteed alpha. The wise approach is to remain focused on fundamentals and long-term objectives.