New data from the Financial Conduct Authority lays bare the AI investment advice risks facing younger UK investors, with almost half of those surveyed wrongly believing that AI-generated financial information is subject to regulatory oversight. The FCA’s survey covered UK adults aged 18 to 40 who own or are considering investments, and the results raise questions that anyone building a long-term portfolio, or advising one, should sit with carefully.
How far trust in AI has outpaced understanding
Among respondents, 56% said they trust AI tools to guide financial decisions, outstripping trust in television and radio (47%), the press (46%), and social media influencers (29%). Four in five less experienced investors have used AI for investing assistance, with two-thirds relying on it occasionally or regularly and expecting to increase usage over the coming year.
Those headline figures alone are instructive for anyone thinking about where retail behaviour is heading. But the misconceptions sitting beneath them are where the real concern lies. Nearly half (44%) mistakenly believe AI-generated financial information is regulated, while 38% consider it acceptable to base investment decisions solely on AI outputs. A further 32% wrongly believe they would be entitled to compensation from the Financial Services Compensation Scheme or the Financial Ombudsman Service if AI-driven advice resulted in losses.
That last figure deserves to be read twice by anyone in, or approaching, drawdown. The FSCS and FOS protections that underpin confidence in regulated financial advice do not extend to general-purpose AI chatbots. The FCA has been explicit: those platforms sit outside its regulatory perimeter entirely.
AI investment advice risks sit in a broader supervisory picture
This survey does not arrive in isolation. On 21 November 2024, the Bank of England and FCA published the results of their third survey on AI and machine learning across UK financial services, according to Global Regulation Tomorrow. That survey incorporated questions on generative AI for the first time, reflecting the technology’s rapid growth since the 2022 edition. The regulators are clearly aware that the landscape has shifted; what the consumer-facing data now shows is that retail understanding has not kept pace.
Lucy Castledine, director of consumer investments at the FCA, acknowledged the legitimate research uses of AI while drawing a clear boundary: ‘AI can help you research companies, understand jargon or explore options before you make a decision. But you need to understand how you’re protected and continue to use your own judgement.’ She directed consumers to the FCA’s InvestSmart website as a cross-checking resource.
That boundary matters because it maps onto the risk framework any conservative investor should already be using. Research assistance is one thing; substituting an unregulated output for regulated advice, particularly on decisions involving pension drawdown, ISA allocation or inheritance planning, is quite another.
Sam Christopher, proposition director at Quilter, framed the issue in terms of trajectory rather than the present moment: ‘As AI gets more sophisticated, and people begin to use it more in their daily lives, then we can expect AI agents will become one source of information used by many in their financial planning and likely in a much more personalised way.’ He was also direct about the hierarchy: ‘Personalised, and regulated, financial advice from a human is still likely to result in the most positive outcome for consumers, especially when there are big decisions to be made.’
Christopher also advised that, where AI tools are used, consumers should provide sufficient context in prompts without divulging sensitive personal information, and should cross-check all outputs against established authorities such as HMRC or GOV.UK before acting.
Rob Hillock, head of personal financial planning at Broadstone, put the sequence-of-returns risk in starker terms. ‘AI is rapidly becoming the first port of call for a new generation of retail investors, but confidence is clearly running ahead of understanding,’ he said. ‘The rapid growth of low-cost trading apps has put stock-picking and crypto investment within easy reach, while AI can appear to offer free, instant guidance on which investments will be the next winners, creating a potentially dangerous combination.’
For investors with a five-to-ten-year horizon or longer, the practical implication is straightforward: AI is a research aid, not a regulated advice channel. Over a longer accumulation phase, the cost of an error compounded by misplaced trust in an unregulated source can be substantial. The FCA’s InvestSmart resource and any FCA-authorised adviser remain the appropriate starting points for decisions that carry real financial consequence.
