AI financial advice is surprisingly good, study finds

Large language models give surprisingly good financial advice, according to a new MIT Sloan study that has drawn wide attention on Hacker News. The research, led by MIT Sloan's Taha Choukhmane and co-authors, found that LLMs encourage smart financial behavior, though they still fall short on the more subtle aspects of saving and investing.

The study suggests that asking the right questions matters as much as the model itself. When prompted carefully, AI assistants delivered advice that improved financial decision-making, from budgeting to debt management. Researchers noted the models perform best on straightforward, well-defined financial questions, where their responses are practical and easy to act on.

The gaps appear in nuanced territory: tax strategy, complex investment trade-offs, and long-term planning where context and personal circumstances play a large role. The authors argue this is not a reason to dismiss AI financial advice, but a signal for how to use it, as a complement to professional guidance rather than a replacement. The findings add to a growing body of research on where LLMs genuinely help, and where they still need human judgment.