The share of Americans using artificial intelligence (AI) for financial tasks surged from 10% to 55% in just one year, according to TD Bank’s 2026 U.S. AI Insights Report.
A Vanguard report published in September 2026, based on a June 2026 survey of 6,686 investors, delivers the clearest picture yet of what happens when consumers trust chatbots with their personal finances.
Among those who used chatbots, 37% reported receiving incorrect or misleading information, and 44% said they could not judge its accuracy.
That leaves only 19% of chatbot users expressing confidence in the quality of the information they received. This aligns with research from the Massachusetts Institute of Technology (MIT), which found structural flaws in how large language models (LLMs) handle personal financial decisions.
Also read: ChatGPT just became the cheapest financial advisor in America
Vanguard’s survey reveals who uses chatbots and who trusts them
About 32% of Vanguard investors reported using chatbots for financial guidance, including personal finance, investing, and retirement planning. Younger investors are more likely to use chatbots for financial advice.
About 43% of Gen Z and Millennial investors have tried them, compared with just 15% of baby boomers and older investors.
More Vanguard:
- Vanguard spotlights Roth IRA trap hiding in your income
- Vanguard’s global ETF flags a stark shift for investors
- Vanguard uncovers the real cost of early Social Security
Confidence in chatbot guidance remains weak even among users who have tried it, and the trust gap between chatbot and human advice is stark.
Vanguard found that across Gen Z, Millennial, and Gen X investors, 63% reported low or no trust in chatbot-generated financial guidance, while fewer than 6% reported high trust. Among boomers and older generations, the share with low or no trust rose to 78%.
Roughly half or more of chatbot users reported using the technology to learn about a topic, explore ideas, or pressure-test a concept, while fewer than 4% used chatbots to make financial decisions on their behalf, the survey showed.
MIT research identifies deeper structural flaw in chatbot financial guidance
Andrew Lo, director of the Laboratory for Financial Engineering at MIT, and Jillian Ross, research affiliate at MIT, tested GPT-4o alongside three GPT-5.4 variants on 1,000 synthetically generated client profiles in an April 2026 paper.
They identified what they called “heuristic collapse,” a systematic pattern in which chatbots base investment recommendations almost entirely on self-reported risk tolerance while functionally ignoring income, debts, time horizon, and other variables.
The distinction matters because the suitability standard governing registered investment advisors requires individualized reasoning based on a client’s full financial circumstances. When Ross and Lo equipped the chatbot with web search access to partially offset this tendency, the heuristic collapse was reduced.
The accuracy problem extends beyond reasoning. Major chatbots hallucinated between 20% and roughly 77% of the finance citations they generated, a 2024 study “Evaluating the Accuracy of Chatbots in Financial Literature,” found, a finding the Vanguard report highlighted.
Chatbots have also been shown to produce biased portfolio recommendations based on a user’s gender, race, and ethnicity, inducing simulated wealth gaps across demographic groups.
Why chatbot adoption is boosting demand for human financial advisors
A regulatory blind spot helps explain why investors are turning back to human advisors, despite growing adoption of chatbots.
Vanguard’s survey found that about 38% of older investors and 16% of Gen Z and Millennial investors either didn’t know whether chatbots have a legal obligation to act in their best financial interests or mistakenly believed that obligation already exists.
The Vanguard survey found results that challenge the idea that chatbots will eventually replace human financial advisors. AI is 11 times more likely to increase the perceived value of human advice than to decrease it, with 32% of investors saying they value human advisors more because of AI, and only 2.9% saying the opposite.
Among clients already working with a human advisor, 41% said AI made them value that relationship more, while only 2% said the reverse.
Lo told CNBC that chatbots lack the legal obligations and personal accountability that make human financial advisors trustworthy.
<strong>What they don’t have is that fiduciary duty. They don’t have the ability to suffer consequences if they make a mistake to the same degree that a human advisor does</strong>.
The accountability gap Lo described helps explain why four in five investors across all age groups said they are uncomfortable with chatbots taking financial actions on their behalf, including 80% of Gen Z and Millennials.
How to use chatbot research without putting your financial plan at risk
Lo has recommended that investors use chatbots for orientation, leveraging them to frame questions, map out available options, and prepare for conversations with a human advisor rather than treating the output as a finished plan.
Among investors surveyed, 46% would be more comfortable with chatbot guidance if a human advisor reviewed the output first, and 74% want some form of financial help from human professionals over the next two to three years.
Lo told CNBC that the greatest risk emerges when investors act on highly specific calculations a chatbot generates about their individual situation without verifying those numbers against a professional who is legally required to account for their full financial profile.
Related: Microsoft CEO wants AI treated like a risky insider at work
