AI Investment Models Favor Young Wealthy Men, Study Shows

A study by the University of California, Berkeley’s Haas School of Business assessed how artificial‑intelligence models emulate investor preferences. Researchers surveyed over 1,200 U.S. adults about their choices among stocks, bonds and cash, then prompted GPT‑4 to role‑play respondents with varying age, gender and income. The AI accurately reproduced well‑known investment patterns – men favored stocks, higher‑income investors leaned toward stocks and bonds, and older participants preferred safer holdings. However, when left unguided, the model defaulted to the stereotypical profile of young, wealthy men, reflecting a bias that could perpetuate inequality in robo‑advisory services.

The authors note that generative AI can be steered toward more personalized advice if explicitly supplied with demographic information, suggesting a path to mitigate bias. They caution that without deliberate design, algorithmic finance tools may embed existing disparities, underscoring the need for inclusive data and safeguards as AI‑driven investment platforms expand.