The role of AI recommendations in extending the Black-Litterman portfolio

Abstract

Purpose – This study explores the role of artificial intelligence (AI) recommendations in portfolio optimization by extending the Black-Litterman (BL) model using consensus analyst opinions generated by ChatGPT. The aim is to assess if AI recommendations can improve portfolio diversification and risk-adjusted returns compared to traditional investment strategies. Design/methodology/approach – We conducted a quantitative analysis using weekly historical price data across equities, commodities, fixed-income securities and cryptocurrencies from January 2018 to May 2023. Portfolios constructed with the extended BL model were tested against standard benchmarks, including the S&P500 index and various mean-variance portfolios. Out-of-sample performance and robustness were evaluated through 100 random resampling procedures. Findings – Results indicate that integrating AI-generated analyst consensus significantly improves the BL portfolio’s risk-adjusted returns. The AI-enhanced model consistently outperformed traditional mean-variance portfolios, the unadjusted BL model and market benchmarks. Robustness tests confirmed the method’s stability and practical feasibility in real-world investing. Practical implications – Portfolio managers and individual investors can apply this enhanced BL model for more effective asset allocation decisions. Using AI-generated recommendations simplifies the integration of broad analyst perspectives, reduces reliance on subjective human judgments and leads to portfolios that deliver stronger and more consistent risk-adjusted performance. Originality/value – This research is the first to integrate ChatGPT-generated analyst recommendations directly into the BL framework. It addresses critical limitations of modern portfolio theory, particularly estimation errors, offering a practical solution leveraging AI advancements for portfolio optimization.