AI in action · Investor protection
Vanguard: 37% of Investors Using AI Got Misleading Answers
In a Vanguard survey of 6,686 of its investors, 37% of those who had used AI for money questions said it gave them incorrect or misleading information. Another 44% could not tell, and eight in ten said they were uncomfortable letting AI act for them.
Chatbots have become a common first stop for financial questions, and a large survey suggests users are noticing the errors. Vanguard's research team polled 6,686 of its own investors in June 2026 and published the results on September 30 in a paper titled The AI Advice Frontier: Use, Trust, and the Human Edge. Roughly one in three respondents had used AI for personal finance, investing or retirement planning, rising to about four in ten among younger investors. Of those users, 37% said the tools had given them incorrect or misleading information, 44% said they could not judge the quality of what they got, and only about one in five was confident the answers were sound.
How people use the tools helps explain the low trust. InvestmentNews, which reported the findings on October 1, noted that about half or more of AI users treat it as a starting point: learning a concept, getting going on a topic or testing an idea. Only 4% of younger investors let it make decisions for them, and respondents were about three times as likely to use AI to prepare for or review a meeting with an adviser as to replace one. The reasons given for using it at all were mundane: it is free, it satisfies curiosity and it is available at any hour.
Trust numbers were stark. Among Gen Z, millennial and Gen X investors, 63% reported little or no trust in conversational AI guidance and only 5% to 6% reported high trust; among baby boomers and older, 78% had little or no trust. By contrast, between 54% and 63% of every age group said they trusted human advisers highly. Between 80% and 85% of each group said they were uncomfortable with AI acting on their behalf, a figure worth setting beside the brokerages now letting outside agents place trades in customer accounts.
The survey found one area where AI scored well. About four in ten users said chatbots were better than human advisers at avoiding conflicts of interest and giving unbiased guidance, and Vanguard's summary was that AI competes on perceived objectivity while people keep the edge on trust and empathy. It also found a knowledge gap with consumer-protection consequences: 37% of older investors did not know, or were unsure, that AI tools are not bound by the fiduciary duty that requires a registered adviser to act in the client's best interest.
Vanguard used the findings to argue for policy changes, which should be read as one firm's position: a clearer line between general education and personalised advice, the same standards for functionally equivalent advice whether it comes from a person, a regulated firm or software, and stronger know-your-customer and data protections before agents handle money. For an individual investor the practical reading is simpler. An answer from a chatbot carries no duty of care, and more than a third of the people who relied on one in this sample found it wrong at least once.
Sources: Vanguard — The future of advice: what AI means for investors, advice and policy (Sep 30, 2026); InvestmentNews — AI financial advice lifts demand for human advisors, says Vanguard (Oct 1, 2026)