Can AI news improve the out-of-sample forecasts of risk premia and volatility in the U.S. stock market?

Abstract The global adoption of AI technology has created notable investment opportunities in the finance sector. This study investigates how AI-related information from online news and web searches predicts risk premia and volatility in AI stocks compared to traditional stocks without AI components. Our findings indicate that AI news significantly lowers the risk premia while increasing return volatility for both stock categories; however, the significant dampening of the risk premium is stronger for AI stocks. This translates to AI stocks bearing a lower risk profile in the face of AI news and thus a lower risk premium, highlighting the notion that AI stocks are perceived as a more stable investment vehicle compared to conventional stocks; this is due to the disruptive role of AI technologies in driving innovations that are seen to be more representative of the future, more valuable and thus less risky. Furthermore, integrating AI news improves forecasting accuracy for both AI and conventional stocks across various timeframes and metrics. Moreover, the forecasting gains from models that incorporate VIX are more pronounced than those from models that incorporate inflation. This research is the first to analyze the distinct effects of AI news on AI stocks versus traditional stocks, making a significant contribution to the field.