Imagine this scenario: Susie, a 63-year-old recent retiree, is weighing her options on when to start receiving Social Security while also managing her retirement savings to minimize her tax burden.
With the help of an AI chatbot, she enters her information and receives organized, confident answers. “Claim now to convert this amount. Here’s why.”
The chatbot presents itself authoritatively, leading Susie to rely solely on its advice, without consulting her financial planner. While the guidance may have been sound, it potentially overlooked critical details like her spouse’s youth and poor health. You can reverse social security calculations. It could also have missed that the proposed retirement plan conversion might lead to unexpected Medicare premium increases in two years.
Susie may never truly know if the guidance she received was right for her. Unlike a human advisor, AI lacks the ability to follow up and inform her of uncertainties.
Susie’s experience isn’t unique. An alarming number of individuals utilize AI for financial advice. According to a 2025 Pew Research Center Study, 34% of U.S. adults and 58% of those under 30 have used chatbots like ChatGPT, with market share doubling in just two years.
More users are consulting AI for financial matters, with some facing steep consequences. A 2025 Survey of 2,000 U.S. Adults revealed that 19% reported losing over $100 due to financial advice from AI chatbots, a statistic that rose to 27% among Gen Z investors.
These aren’t hypothetical risks; people make financial decisions based on AI’s outputs, which may lead to significant losses.
Get the world’s most fascinating discoveries delivered straight to your inbox.
As a finance professor, I closely monitor AI’s integration into personal finance, and this aspect of the technology deeply concerns me.
The Discussion Around AI is Misguided
Two conflicting observations regarding AI exist. On one hand, users overly trust chatbots, treating them like infallible sources. This phenomenon has been documented by researchers. Algorithm evaluation. Conversely, many individuals lack enough trust in AI, leading to missing opportunities.
I contend these perspectives are interconnected. Whether you view AI positively or negatively depends on its ability to recognize when it is incorrect.
When AI makes evident errors, it tends to lose credibility, prompting users to seek expert opinion sooner. This situation becomes a safe fail.
On the other hand, a dangerous failure occurs when AI delivers incorrect yet confident answers. Without the ability to identify these mistakes, users may persist in relying on the chatbot, delaying the search for professional help.
Financial matters commonly encounter the latter type of failure, leading to detrimental consequences.
Younger users, particularly men, are the primary demographic seeking financial advice from chatbots.
(Image source: Krongkaew via Getty Images)
When Fluency is Misinterpreted as Accuracy
Three factors make financial advice from AI particularly risky.
First, fluency is not synonymous with accuracy. Individuals often perceive articulate, confident responses as competent. Yet, the sophistication of an answer does not guarantee its relevance or accuracy. A chatbot may articulate tax-related information flawlessly, but it may still miss critical details specific to users.
Second, AI tools are the least reliable where risks are greatest. They perform well in everyday topics, such as general financial concepts like Roth IRAs and compound interest.
However, financial decisions often involve rare, intricate, unique scenarios such as exercising stock options, navigating alternative minimum tax, or drafting divorce settlements.
I previously discussed the implications of AI trading on Wall Street, where rare market events yield little data for AI training, leading to AI’s most confident but least accurate recommendations.
Concerns persist regarding AI trading bots creating new financial risks similarly to personal finance.
This discrepancy in reliability is termed the “Jagged Frontier”—trustworthiness in routine scenarios but unreliability in complex cases, which often lead to significant financial losses.
Third, the consequences of financial advice are often difficult to observe. Economic advice falls under the category of “credence goods”, much like a mechanic’s diagnosis or a doctor’s prescription. Users might remain unaware of the advice’s accuracy until years later. Tax mistakes might not be uncovered until an audit occurs.
401(k) withdrawal strategies may not reveal issues until market downturns. Without immediate feedback, erroneous but confident responses remain unchecked.
Therefore, the above statistics from Pearl likely underrepresent the issue, as they only reflect the losses individuals have noticed.
The Quiet Failures Are the Most Concerning
In Susie’s situation, the most significant risk wasn’t one major error. The chatbot’s confident advice discouraged her from consulting with a professional, which could have been beneficial.
The true danger lies not in following poor advice but in the absence of seeking sound guidance. As technology becomes smoother and more comforting, users may stay in DIY mode longer, even when professional advice is essential.
Who is most vulnerable? A study on robo-advisors in India involving co-author Vishal Bawkarkaran identified that users skews younger, are predominantly male, and usually consist of small retail investors and professionals. New account registrations spiked during volatile market periods.
Thus, those primarily relying on automated advice also align with the 27% of Gen Z individuals who reported losing $100 or more from chatbot consultations during market turmoil.
Furthermore, various incentives are worth noting. In my recent analysis, I contend that platforms profiting from user engagement may prioritize sounding useful and confident, potentially retaining users who would have been better served by human advisors.
A system designed for engagement does not equate to one tailored for safeguarding your financial future. The tension between these priorities poses a risk. Financial industries, as noted by Bloomberg, are already experiencing a crisis, driven partially by AI-driven disruptors. Recently, a new AI tax tool caused wealth management stocks to decline, sparking investor fears.
Making Wise Choices with AI
These insights don’t imply you should avoid AI for financial guidance. When used appropriately, these tools can serve as valuable resources for financial education.
However, keep in mind that financial advisors aren’t always right. Conducting thorough research to ensure they meet certain established criteria as outlined by the Consumer Financial Protection Bureau is crucial. Additionally, price transparency is essential.
When utilizing AI, knowing when to draw the line is key.
Consider AI as a tool for initiating discussions rather than making decisions. It’s excellent for learning concepts and formulating questions before consulting experts. AI can help users refine their vocabulary for deeper conversations.
Stay vigilant for signs that indicate you’ve ventured into areas where AI is less reliable. Red flags include significant financial stakes, tax ramifications, irreversible choices, and specifics related to individual situations rather than general rules.
Real estate matters, retirement withdrawals, Social Security claiming strategies, business organization, and major one-time transactions require human expertise, potentially from a certified financial planner.
Moreover, remember that confidence doesn’t equate to competence. A seemingly sophisticated and assured answer shouldn’t lull you into complacency. For the trickiest financial inquiries, that smooth confidence signals it’s time to consult an expert.
This editorial article has been republished from The Conversation under Creative Commons License. Please read the original article.
Source: www.livescience.com