Financial advisers’ clients are more likely to trust unchecked AI guidance, study finds

Research suggests investors with traditional advisers may be nearly three times more likely to act on chatbot recommendations without verifying them, raising fresh concerns about misplaced trust in AI-powered money management.

A new study from Evergreen.ai has added a sharper warning to the rush towards AI-powered money management: people who already use a traditional financial adviser appear more likely than others to act on unverified chatbot guidance. The research, published on 16 October 2026, found that 41.9% of advised AI users made financial decisions based solely on chatbot recommendations without checking them elsewhere, a rate the company said was nearly triple that of investors without an adviser.

The result cuts against the assumption that professional advice makes investors more cautious, not less. Instead, it suggests that having a human adviser may create a misplaced sense of security, or encourage people to treat AI as a quick second opinion rather than a tool that needs scrutiny. That is a risky habit in a field where small errors can quickly become expensive mistakes.

The broader concern is that generative AI can sound authoritative even when it is wrong. As Kiplinger has noted in recent reporting on retirement advice, chatbots can produce outdated, incomplete or misleading answers because they do not reliably understand current rules, live market conditions or the personal context that shapes financial decisions. A confident tone does not make the output accurate, and in finance that distinction matters.

The possible consequences are straightforward and serious. Bad AI guidance can lead to investment losses, tax penalties, higher debt costs, missed opportunities and avoidable stress. It can also create a false impression that a plan has been verified when it has not. In practice, that means a user may accept a stock recommendation, tax suggestion or retirement strategy that sounds plausible but has not been checked against authoritative sources.

Kiplinger’s guidance on fact-checking AI retirement advice points to a basic discipline that is increasingly relevant well beyond retirement planning: verify dates, cross-check figures with the relevant authorities and involve a qualified human expert before acting on anything irreversible. That advice is especially important because many chatbots were trained on information that may no longer reflect current tax rules, savings thresholds or regulatory guidance.

The distinction between general-purpose chatbots and regulated financial tools is also important. Robo-advisers such as Betterment or Wealthfront are built to manage portfolios within defined rules and oversight structures. By contrast, a chatbot asked a broad finance question may generate polished text without any obligation to disclose uncertainty, explain its reasoning properly or protect the user from a poor decision. Kiplinger’s testing of ChatGPT, Claude and Gemini on personal finance scenarios found that the tools could be useful for basic guidance, but often lacked the follow-up questions and contextual judgement that a skilled adviser provides.

That is why the latest findings should be read as both a warning and a reminder of what human advisers still do better. They can account for family circumstances, behavioural biases, tax complexity and emotional pressure during market swings. They can also challenge assumptions, correct misunderstandings and take responsibility for guidance in a way AI cannot. As AI tools become more common in personal finance, the safest approach is not blind trust but disciplined checking, with AI used for exploration and a qualified human adviser retained for decisions that matter.