MIT, Stanford Study Finds AI Financial Advice Improves 5% With Better Prompts, Shows Gender Bias
Updated
Updated · USA TODAY · May 15
MIT, Stanford Study Finds AI Financial Advice Improves 5% With Better Prompts, Shows Gender Bias
4 articles · Updated · USA TODAY · May 15
1,000 Americans’ draft chatbot questions formed the basis of an MIT-Stanford study that found AI usually gives sound basic money advice but misses investing nuances such as rebalancing and Consumption smoothing.
Roughly 5% higher long-run returns emerged when the same chatbot was prompted with questions written by people with higher financial literacy, showing answer quality depends heavily on prompt quality and user experience.
55% of Americans now use AI for financial help, up from 10% in 2025 in a TD Bank survey, raising the stakes as chatbots become more common than human advisers for many users.
Women received more conservative recommendations—less stock exposure and safer allocations—and some of that gap remained even when researchers posed the same question while signaling different genders.
A separate Harvard-Boston College study found 13% of Seeking Alpha articles were AI-assisted before a 2023 ban, and those pieces drew less engagement and weaker market impact than human-written analysis.
Is AI financial advice democratizing wealth-building, or creating new risks for the financially vulnerable?
AI reflects gender bias in financial advice. What other societal biases are coded into the algorithms managing our money?
When an AI advisor gives flawed financial advice, who is legally and financially responsible for an investor's losses?
AI in Financial Advice: MIT-Stanford Study Reveals 1,000-User Test, Gender Bias, and Urgent Need for Oversight
Overview
A major MIT-Stanford study in May 2026 explored how artificial intelligence is changing financial advice. By analyzing AI-generated advice for 1,000 Americans, the study found that the quality of recommendations depends heavily on how clearly users describe their needs and goals—better questions lead to better answers. However, many people struggle to ask precise questions, which can limit the usefulness of AI advice. The research also highlighted that AI systems can show gender bias in their recommendations, raising concerns about fairness and the need for better oversight and more inclusive training data in financial AI tools.