AI chatbots may need regulatory oversight, FCA warns
More than a quarter of Britons seek financial guidance from AI
Financial Conduct Authority (FCA) Executive Director Sheldon Mills said the regulator should examine the consequences of general-purpose AI models looking like regulated financial guidance.
Britain should consider whether popular AI chatbots such as ChatGPT, Gemini and Claude ought to fall under financial regulation as more consumers rely on them for money-related advice, according to a senior official at the UK's financial watchdog.
Financial Conduct Authority (FCA) Executive Director Sheldon Mills said the regulator should examine whether general-purpose AI models have evolved into services that resemble regulated financial guidance.
His comments come ahead of the publication of an FCA-commissioned review into the impact of AI on financial services, which is due to be considered by the regulator's board.
The review found that more than a quarter of UK consumers already trust AI assistants, including Anthropic's Claude, OpenAI's ChatGPT and Google's Gemini for financial advice. It also found that one in five adults would be willing to allow AI systems to make financial decisions on their behalf, such as choosing savings products or borrowing options. That is despite those services not being covered by the same consumer protections that apply to regulated financial advisers.
Mills said this raises questions over whether conversational AI systems were beginning to perform functions that were economically similar to regulated financial recommendations.
New forms of systemic risk
The review warns that AI's rapid adoption could create new forms of systemic risk.
Around 81% of financial firms worldwide are now using AI to some degree, with roughly 40% deploying it at scale or as part of broader business transformation, according to industry research cited in the report.
While much of that use remains focused on back-office operations, UK firms are increasingly introducing AI into customer-facing services, including complaint handling and investment guidance.
Mills said regulators would need to adopt AI themselves to keep pace with the technology's rapid development.
"It is an arms race," he said, arguing that AI could help regulators identify emerging risks, detect misconduct and strengthen oversight.
The report also highlights concerns that banks and other financial institutions could become overly dependent on a small number of technology companies supplying AI models, cloud computing and other critical infrastructure.
Such concentration, it says, could increase the likelihood of common failures or encourage firms to behave in similar ways during periods of market stress, potentially amplifying risks across the financial system.
Threats and opportunities
The review identifies both threats and opportunities from wider AI adoption. It says AI could improve access to sophisticated financial guidance by offering personalised support to consumers who would otherwise be unable to afford professional advice.
Mills argued that technology had the potential to "democratise" financial services by making expertise previously available only to wealthy clients accessible to people on more modest incomes.
However, the report also warns that hyper-personalised AI systems could introduce bias, enable opaque pricing practices and make consumers more vulnerable to manipulation.
It recommends that the FCA work with government and industry to develop an AI-powered public financial guidance service offering free information to consumers.
The review also cautions that advances in AI are likely to increase fraud and cyber threats through deepfakes, synthetic identities and increasingly sophisticated social engineering attacks, while recommending that the technology itself should also be used to strengthen cyber defences.
The FCA's board is expected to consider the report before deciding whether to act on its recommendations.