In this episode, the Concretum Research Team discusses a new study at the intersection of artificial intelligence and systematic investing. Conducted with Finreon Asset Management, the research asks a simple but ambitious question: can a large language model like ChatGPT read corporate news and use that information to refine momentum-based investment strategies?

Using daily data on S&P 500 stocks, the study combines traditional price-based signals with ChatGPT’s interpretation of firm-specific news. The results show that this AI-assisted approach improves both stock selection and portfolio performance, delivering higher risk-adjusted returns than a standard momentum strategy — even in out-of-sample tests.