AI is rapidly transforming the world of personal investment. Retail savers around the world are asking chatbots for investment strategies, debt management plans and stock tips. Adoption is happening quickly: research found a fifth of UK adults were open to using AI models to make financial decisions for them.
UK regulators are taking notice. The Financial Conduct Authority this week published the Mills Review, which highlights new risks from the use of AI in retail financial services.
Regulators are wise to pay attention to emerging technologies but should avoid saddling them with excessive regulations. As governments around the world develop responses to the use of AI in retail financial services, they can mitigate risks while ensuring that citizens can still reap the benefits of large language models.
AI tools could empower a generation of savers to take control of their financial futures and invest their money wisely. LLMs are increasingly filling the role that personal finance books and blogs have played, educating everyday people on ways to invest their money. Beyond summarising points, a model can tailor its responses to a set of criteria, answer specific questions and redirect users to more information. The technology has the potential to democratise access to useful resources.
But these technologies also raise new questions. In the UK, companies that provide investment advice have registration requirements, fiduciary responsibilities and disclosure duties. The law distinguishes between formal advice, provided by a regulated adviser and tailored to an individual’s circumstances, and generic guidance, such as a static webpage that discusses investment strategies. This distinction can be blurred by the emergence of generative and agentic AI financial tools.
There is also the problem of liability. Customers have recourse against regulated financial advisers who give poor advice, but AI agents are information aggregators that do not provide users with the same liability protections. The Mills Report highlights how many retail financial services customers do not appreciate this distinction: only 40 per cent of survey participants understood that they had no formal recourse against bad financial advice from general-purpose AI.
Small interventions could play a role in helping consumers avoid financial misfortune. AI models should provide disclaimers when sharing anything that could be misconstrued as investment advice. And as services such as OpenAI’s ChatGPT roll out advertisements, in the UK they will be required to comply with disclosure requirements for sponsored financial content.
The FCA should avoid being overzealous in its approach to regulation. Saddling companies with excessive burdens that aim to protect consumers could in fact do them harm. In the past, the FCA’s Retail Distribution Review attempted to improve the quality of investment advice, but instead increased the cost of advice for the average consumer, leading to an “advice gap”. Over-regulation of financial advice pushed many people towards online “finfluencers” who offer lower-quality information. Regulators might learn from past mistakes as they consider regulating AI.
New financial technologies highlight the urgent need for widespread personal finance education, including as part of school curricula. Young people ought to be taught traditional financial literacy skills and educated on the uses and limitations of AI tools for their finances. Regulation will always lag behind the latest technology. An educated public is the best protection against any harm from AI advice.