MIT Study: AI Financial Advice Is Surprisingly Effective
Original: AI financial advice is surprisingly good, especially if you ask right questions
Why This Matters
AI is now a primary financial advisor for half of Americans, making advice quality a systemic concern.
MIT Sloan research finds that following AI financial advice can build significant savings buffers for adults over 30. AI consistently recommends saving during working years, diversified stock investment, and reduced equity exposure after 45 — but struggles with shocks like unemployment and portfolio rebalancing.
A new study from MIT Sloan School of Management, co-authored by finance professor Taha Choukhmane, evaluated the quality of financial advice generated by large language models (LLMs). The research found that roughly half of Americans now use AI for financial guidance, yet little was previously known about the advice quality or whether users act on it.
Key findings show AI consistently encourages saving during working years, drawing down savings in retirement, heavy investment in diversified stock funds, and reducing stock exposure after age 45. Following these recommendations can create meaningful savings buffers for virtually all individuals over 30.
However, the study identified notable limitations: AI chatbots struggled to adapt advice following economic shocks such as unemployment, and they tended to allow portfolios to drift rather than actively rebalancing them. Structured prompts improved advice quality, but insufficient rebalancing guidance remained a persistent issue. The research also found variation in advice quality based on the user's gender, financial literacy, and familiarity with LLMs.