AI advice and the retail investor: FCA data reframes the suitability problem

The FCA has put a number on what boards have suspected for months. Four in five less experienced investors aged 18 to 40 have used AI for help with investing, and around two-thirds do so occasionally or regularly (FCA). More striking for anyone running a retail investment business: 44% of these investors mistakenly believe AI-generated financial information is regulated, and 32% think they would get compensation from the FSCS or Financial Ombudsman Service if AI advice went wrong (FCA).

Key Executive Takeaways

• FCA research published on 27 August 2026 shows most young UK investors now use general-purpose AI tools for investment research, and a substantial minority wrongly assume those tools sit inside the regulatory perimeter.
• Firms serving retail investors face a widening gap between where clients form investment views (unregulated chatbots) and where liability and suitability obligations actually sit (the authorised firm executing the trade).
• Senior leaders in wealth, platforms and advice should assume the FCA will treat AI-driven consumer misunderstanding as a Consumer Duty issue, not a future problem, and adjust disclosures, onboarding and vulnerability frameworks now.

The trust transfer

The headline finding, that 56% of these investors trust AI tools, more than TV and radio (47%), press (46%) or influencers (29%), is less interesting than the direction of travel (FCA). Two-thirds expect to lean on AI more over the next year. For incumbent wealth managers and platforms, the competitive question is no longer whether a robo-adviser will disintermediate them. It is whether a general-purpose chatbot, sitting entirely outside the FCA's remit, is already shaping the investment thesis before the client ever logs in. Lucy Castledine, director of consumer investments at the FCA, framed AI as something that can help investors 'research companies, understand jargon or explore options before you make a decision' but stressed that 'the final call is yours' (FCA). That is a supervisory position as much as consumer guidance.

Where the liability actually lands

The FCA is explicit that general-purpose AI chatbots are not regulated, though tools set up specifically to provide financial advice would likely fall within its remit (FCA). That perimeter line is clean in law and messy in practice. If 38% of young investors think it is acceptable to make an investment decision based solely on AI outputs, the authorised firm receiving that order is executing against a decision process it cannot see and did not shape (FCA). Under the Consumer Duty, the question of whether a client understood what they were buying, and whether the firm did enough to test that understanding, becomes harder to answer with a straight face. Appropriateness assessments designed for a pre-generative-AI world will not survive contact with a client base that treats ChatGPT as a research analyst.

What boards should be asking

Three questions belong on the next risk committee agenda. First, does the firm's target market analysis reflect the reality that a majority of younger clients are arriving with AI-shaped views. Second, do execution-only disclosures still hold up when a third of clients wrongly expect FSCS cover for AI-influenced decisions. Third, how will the firm evidence, to a supervisor, that it identified and mitigated foreseeable harm from AI misuse when the FCA has now put the data in the public domain. The 86% of investors who understand the need to check sources is reassuring on paper (FCA). It will not be a defence.

The FCA has moved the AI conversation from model risk inside firms to consumer behaviour outside them. Boards that treated retail AI as a marketing question now have a supervisory one.