AI Chatbots Deliver Inconsistent and Occasionally Biased Financial Advice

A study published in the Journal of Financial Planning tested seven widely used generative‑AI tools—including ChatGPT, Claude, Copilot, DeepSeek, Gemini, Meta AI and Perplexity—by presenting identical household‑finance scenarios. The researchers found the systems gave divergent recommendations on emergency‑savings levels, sustainable retirement‑withdrawal rates and stock‑bond portfolio mixes. In several cases the advice shifted when only the user’s race or gender was altered, indicating potential bias.

Consumer reliance on AI for money matters is growing. A U.S. survey cited by the study shows 19 % of adults have used AI tools for personal‑finance queries, with higher usage among younger generations. An associated poll reported that roughly one‑fifth of respondents lost more than $100 after following AI‑generated advice, a figure that rose to 27 % for Generation‑Z investors. In the United Kingdom, a Zable survey found 76 % of Britons would not verify a financial adviser’s credentials before acting on advice, and 29 % reported monetary losses after bad guidance; testing of AI tools showed most defaulted to U.S.–centric recommendations.

Regulators have taken note. Britain’s Financial Conduct Authority warned that unregulated AI advice can blur the line between general information and regulated financial counselling, exposing consumers to harm without the protections afforded by licensed advisers. The study authors urged continued scrutiny to ensure transparency, fairness and accuracy in AI‑driven financial guidance.