JPMorgan tests AI agents for asset management instead of humans
7/10/2026, 02:39 PM • Евгения Слив

Shares of the American banking giant JPMorgan Chase confidently rose against the backdrop of sensational news that investment agents based on artificial intelligence managed to outperform the traditional portfolio with a classic sixty-forty allocation during massive historical simulations. Bank researchers developed a unique system of AI agents that dynamically reallocates assets between stocks and bonds, instantly reacting to changing market conditions. According to the authoritative publication Bloomberg, during backtests covering the last two decades, the most effective neural network model outperformed the classic portfolio by zero point seven percent annually, while also demonstrating lower volatility and confidently beating JPMorgan's own rules-based market regime model.
Nevertheless, the bank's management urges investors to maintain healthy skepticism, emphasizing that the obtained results are based exclusively on historical simulations, not on real capital injections. Strategists led by Thomas Salopek note that this work should be viewed as the company's first attempt to create an AI system for determining market regimes, rather than as indisputable proof of machines' ability to consistently beat the markets. Experts strongly warn against the uncritical perception of in-sample and overly confident responses from artificial intelligence, asserting that agent AI must rely on a well-thought-out human process of asset allocation, rather than naively assuming that the algorithm itself is the absolute source of financial wisdom.
This technological breakthrough fits perfectly into the global trend of the last two years, within which the world's largest banks are actively integrating large language models into their daily research activities, programming processes, and the creation of internal investment tools. Now, financial institutions are moving to the next, much more complex stage, testing whether these advanced systems can make autonomous and balanced decisions regarding capital allocation across various highly competitive markets. Although the technology demonstrates impressive potential for optimizing returns and managing risks, the necessity of human oversight and the consideration of fundamental macroeconomic factors remain an indispensable condition for the safe future of automated finance.
***
The material is prepared solely for informational purposes and does not constitute a financial advice or recommendation.
