Don't trust AI advice on personal finances: study
A study found that artificial intelligence is not convenient for receiving financial advice: we tell you the reasons they point out
Artificial intelligence has become a common tool to resolve questions about savings, investments or retirement. However, a new academic study warns that it is not advisable to completely rely on their answers when it comes to making money decisions, as they may be inaccurate, incomplete, or vary depending on the platform used.
The research, published in the Journal of Financial Planning, analyzed the free versions of ChatGPT, Claude, Copilot, DeepSeek, Gemini, Meta AI and Perplexity. The researchers found significant differences in recommendations related to emergency funds, investment allocations, and retirement savings withdrawals.
“Answers generated by generative AI may appear safe, but may still be incomplete, misleading, or incorrect,” wrote the professors at the University of Georgia and the University of Rome Tor Vergata. Furthermore, they stated that these results raise doubts “about the consistency and impartiality of recommendations based on generative AI.”
As part of the study, the researchers posed the same financial scenarios to all seven tools and then modified only user characteristics, such as race and gender.
The results showed relevant differences between recommendations on how much to save for emergencies and how to distribute an investment portfolio. Although several responses agreed with general financial planning principles, not all of them offered the same suggestions.
“Generative AI can serve as a useful starting point for consumers, but it should complement, not replace, professional financial advice,” the researchers concluded.
The study comes at a time when more Americans are turning to artificial intelligence to manage their finances. According to a survey by Intuit Credit Karma, 66% of those who have used generative AI used it for financial advice. Among members of Generation Z and millennials, the figure reaches 82%.
Experts agree that these tools are useful for explaining general concepts, such as the importance of diversifying investments, but they have limitations when it comes to responding to personal situations.
One of the main risks is that artificial intelligence can generate incorrect answers with a completely confident tone.
“One of the things that particularly worries me about language models is that no matter what you ask them, they will always give an answer that sounds authoritative, even if it isn't,” said Andrew Lo, director of the Financial Engineering Laboratory at the Massachusetts Institute of Technology (MIT). “When it comes to very, very specific calculations about your personal situation, that's where you have to be very, very careful.”
Additionally, small differences in the way a question is worded can change the answer. Specialists remember that these platforms also have no legal obligation to act for the benefit of the user. That is why they recommend using them as an initial guide and always verifying the information before making decisions that may affect personal finances.