“It is an arms race,” Mills said in an interview with the Financial Times. “Some firms have said to us that they feel that this could be an economically equivalent type of service that isn’t regulated and sits outside of the regulatory perimeter.”
A central issue for the regulator is whether conversational responses from AI models should be treated as financial recommendations or guidance under existing rules. While regulated financial firms must comply with strict requirements when providing investment advice, comparable interactions with AI chatbots are not currently subject to the same oversight.
“Is the fact that the chat model might be able to respond to prompts and have a conversation something closer to a recommendation, or guidance?” Mills said.
The report recommends that the FCA carry out a review within the next three to six months to assess consumer risks posed by AI-powered financial services operating outside its remit. That assessment would consider whether the regulatory perimeter should be expanded to cover general-purpose AI systems used for financial guidance.
Alongside potential consumer benefits, the report highlights several risks associated with wider AI adoption. The FCA said AI could make financial advice more accessible by extending sophisticated guidance to consumers who have traditionally lacked access to those services. Mills pointed to the possibility of someone earning £20,000 a year receiving advice typically available only to individuals with £10 million in savings or assets.
“I mean what’s not to like about that?” he said.
At the same time, the report warns that AI could introduce new forms of algorithmic bias, opaque pricing and “personalized manipulation.” It also identifies deepfakes, synthetic identities and targeted social engineering as emerging fraud and cybersecurity challenges.
According to research cited in the report, 81% of financial firms globally are adopting AI in some capacity, with 40% already in more advanced stages of deployment. While many applications remain focused on back-office operations, the FCA noted that firms are increasingly using AI in customer-facing roles, including complaint handling and investment guidance.
The regulator also raised concerns about growing dependence on a small number of technology providers, warning that concentrated reliance on shared infrastructure could create “correlated behaviour, herding and common points of failure across the financial system.”
To address those risks, the report recommends expanding the FCA’s authority under the UK’s “critical third parties” regime, which allows regulators to supervise technology providers that support the financial sector. Companies that could fall within that framework include Anthropic, OpenAI, Amazon, Google and Microsoft. The report says designated companies could face annual self-assessments, additional disclosure obligations and scenario testing for major operational disruptions, although the UK government has not yet determined which companies will receive that designation.
The report also proposes using the UK’s “designated activities regime” to regulate specific AI-related financial activities without requiring every provider to become a fully authorized financial institution.
FCA Chair Ashley Alder said, “We need to keep pace with a rapidly changing environment and the principles-based, outcomes-focused approach we’ve taken on AI.”
As financial institutions begin testing AI agents capable of executing transactions, Mills said accountability should remain with people rather than autonomous systems.
“You need a human on the hook for what they’re doing,” he said.
The report also recommends that the FCA work with public- and private-sector organizations to develop an “AI-enabled financial capability service” that would provide free financial information and guidance to consumers.
Separately, the FCA continues to face scrutiny over its own use of AI after signing a 12-week contract with Palantir to test AI systems for financial crime detection. Some UK lawmakers have questioned whether the arrangement could expose sensitive financial data to US authorities, an allegation both the FCA and Palantir have denied. Mills declined to comment on the contract.
The FCA board is expected to review the report before deciding whether to adopt its recommendations, including possible changes to how AI-powered financial guidance is regulated in the UK.
This analysis is based on reporting from BigGo Finance.
Image courtesy of PYMNTS.
This article was generated with AI assistance and reviewed for accuracy and quality.
