TL;DR: Popular AI chatbots are providing dangerously inaccurate financial advice to UK consumers, with Which? research uncovering misleading tax guidance, incorrect insurance recommendations, and advice that could breach HMRC rules. Up to half of UK consumers may be using AI for financial decisions without regulatory protection.
Widespread Inaccuracy in Financial Guidance
Consumer organisation Which? has revealed concerning inaccuracies across major AI chatbots when tested with 40 financial questions. Microsoft’s Copilot and ChatGPT advised breaking HMRC investment limits on ISAs, whilst ChatGPT incorrectly stated travel insurance was mandatory for most EU visits. Meta’s AI provided wrong information about flight delay compensation claims.
The research found Meta’s AI performed worst, followed by ChatGPT, with Microsoft Copilot and Google’s Gemini scoring marginally better. Perplexity, an AI specialising in search, achieved the highest accuracy score. Estimates suggest between one in six and as many as half of UK consumers now use AI for financial advice.
Real-World Consequences for Consumers
Users report mixed experiences. Kathryn Boyd, a 65-year-old fashion business owner in Ireland, encountered outdated tax codes when using ChatGPT for self-employed tax advice. “It just gave me all the wrong information,” Boyd said, requiring at least three corrections. “Other people asking the same question may easily have relied on the assumptions used by ChatGPT which were just plain wrong.”
Particularly concerning was the chatbots’ handling of tax refund queries. Both ChatGPT and Perplexity promoted premium tax-refund companies alongside the free HMRC service—companies “notorious for charging high fees and adding on spurious charges,” according to Which?
Regulatory Gap Raises Concerns
The Financial Conduct Authority has clarified that advice from general-purpose AI tools lacks consumer protection. “Unlike regulated advice provided by authorised firms, any advice provided by these general-purpose AI tools are not covered by the Financial Ombudsman Service and the Financial Services Compensation Scheme,” the regulator stated.
Tech companies have acknowledged the limitations. Microsoft encourages users to verify AI content accuracy, whilst OpenAI points to industry-wide efforts to improve accuracy, noting their GPT-5 model as “the smartest and most accurate we’ve built.” Google emphasises transparency about generative AI limitations, with Gemini reminding users to consult professionals for financial, legal and medical matters.
Source: The Guardian