A new study of seven generative AI platforms found that financial advice from artificial intelligence can be inaccurate, demographically biased, and vary sharply depending on which tool a consumer uses. The findings arrive as two out of three Americans who have used generative AI say they've turned to it for financial guidance, raising the stakes for both consumers and financial professionals.
Widespread use, uneven advice
The research, published in the Journal of Financial Planning, examined free-access versions of ChatGPT, Claude, Copilot, DeepSeek, Gemini, Meta AI, and Perplexity. The authors-Swarn Chatterjee, Brenda Cude, and Gianni Nicolini-found "significant variation" in how the platforms answered prompts about emergency savings, asset allocation, and retirement withdrawals.
"GenAI-driven responses may sound confident but can still be incomplete, misleading, or incorrect," the paper states. The outputs were often suboptimal or biased, raising questions "about the consistency and fairness of GenAI-driven recommendations."
An Intuit Credit Karma survey published in September found that 66% of Americans who have used generative AI have used it for financial advice. That share jumps to 82% for both Gen Z and millennials.
How the study tested AI platforms
The researchers fed the seven tools identical financial scenarios in August 2025, covering how much to save for emergencies, the optimal withdrawal rate from retirement savings, and recommended portfolio composition. They then repeated the prompts while changing the race and gender of the hypothetical individual to see if the advice shifted.
The results showed "substantial variation in guidance" across platforms. While the tools often echoed generic planning principles-such as the 4% retirement withdrawal rule-suggested emergency savings levels and portfolio allocations differed significantly between them.
The limits of AI for financial planning
Experts caution that AI's confident tone can mask errors. Andrew Lo, director of MIT's Laboratory for Financial Engineering, told CNBC in March: "One of the things about LLMs that I find particularly concerning is that no matter what you ask it, it'll always come back with an answer that sounds authoritative, even if it's not."
Lo added that for highly specific personal calculations, users must be "very, very careful." AI is also sensitive to how prompts are worded, and it has no fiduciary duty to act in a user's best interest. A separate 2024 study in the Journal of Risk and Financial Management found ChatGPT's recommendations to be "generic" and often missing pertinent details, though it could serve as a "first stop" for households seeking advice.
The latest study's authors concluded that GenAI "may serve as a helpful starting point for consumers but should complement, not replace, professional financial advice." They noted that the tools are still developing, and paid versions may produce different results than the free ones tested.
Why this matters for finance professionals
Clients are already using AI to shape their financial decisions. Advisors should expect to encounter plans built on conflicting or biased AI-generated advice and be prepared to correct misconceptions. Understanding the limitations of these tools-and being able to explain them to clients-is becoming a necessary skill. For advisors, staying informed about these tools' capabilities and pitfalls is essential, and targeted AI for Finance training can provide that grounding.
Your membership also unlocks:
