The FCA's Mills Review, published 6 July 2026, found 20% of UK adults, around 11 million people, are likely to use AI that can act autonomously on their finances, while over 75% of UK financial firms already use AI. The regulator will decide within 3 to 6 months whether AI chatbots giving financial guidance should require FCA authorisation.
The Financial Conduct Authority has published The Mills Review, a wide-ranging assessment of how artificial intelligence could reshape UK retail financial services by 2030 and beyond. Commissioned by the FCA Board in January 2026 and led by executive director Sheldon Mills, it is, according to the regulator, the first review of its kind commissioned by any financial regulator globally, and it lands at a moment when consumer and firm adoption of AI in finance is already well underway rather than merely theoretical.
KEY FACTS
- The Mills Review is the first review of its kind commissioned by a financial regulator anywhere in the world.
- Led by FCA executive director Sheldon Mills, commissioned by the FCA Board in January 2026.
- 20% of UK adults, roughly 11 million people, say they would likely use agentic AI acting autonomously within pre-set goals on their personal finances.
- Over 75% of UK financial services firms are already using AI in some form.
- The FCA will review within 3 to 6 months whether large language models such as ChatGPT, Claude and Gemini offering financial guidance should fall under its regulatory remit.
- The review makes 7 priority recommendations to the FCA Board.
Why the FCA commissioned this review
The review builds on existing FCA work including its AI Discussion Paper, AI Sprint, and AI Lab, which includes AI Live Testing and a Supercharged Sandbox developed with NVIDIA. In launching the review in January 2026, the FCA noted that millions of consumers already use generative AI tools to navigate their financial lives and that more than 75% of UK financial services firms are now using AI in some capacity. The regulator described this as a potential inflection point, where advanced, multimodal and agentic AI systems could reshape how financial products are designed, distributed and consumed, potentially including AI systems that make and execute financial decisions autonomously within parameters set by the consumer.
What the review found
The Mills Review identifies four major AI-driven shifts likely to affect UK retail financial services: the transformation of how firms operate internally; the evolution of consumer journeys, as people increasingly interact with AI rather than human staff or static websites; a reshaping of competition and market power, particularly given the small number of large technology companies that provide the underlying AI models; and an amplification of fraud and cyber security risks as AI tools become more capable and more widely available to bad actors as well as legitimate firms.
AI adoption in UK retail financial services
Source: The Mills Review, FCA, 6 July 2026 (Yonder Consulting survey of 5,000+ UK consumers, April 2026). |
As part of the review, Yonder Consulting surveyed more than 5,000 UK retail financial services consumers in April 2026, people holding a day-to-day bank account such as a current or savings account, with quotas set to be representative across age, gender, ethnicity, region, housing tenure and internet ability. The survey found 20% of respondents, equivalent to roughly 11 million UK adults, said they would likely use AI capable of acting autonomously within pre-set goals on their behalf. At the same time, the same consumers expressed concern about trust and control, wanting clear protection when things go wrong and the ability to reach a human being when needed.
The seven recommendations
The review sets out seven priority recommendations for the FCA Board and Executive to consider. These are: securing and adapting the regulatory perimeter; strengthening system-wide coordination and oversight between regulators; monitoring the transition to autonomous AI models and adapting regulatory frameworks accordingly; scaling up the FCA's AI Lab to support AI model and system innovation; enabling the foundations for what the review calls agentic finance; building and adopting an AI-enabled supervisory model for the regulator's own operations; and developing a trusted, public-interest AI-enabled financial capability service.
That final recommendation is notable in its own right: the review suggests the FCA should work with government, industry, consumer groups and the Money and Pensions Service to explore a free, publicly accessible AI-enabled financial support service, on the basis that access to high-quality financial guidance should not depend on a consumer's ability to pay for premium AI tools.
Should AI chatbots be regulated like financial advisers?
Perhaps the most consequential single point in the review is a direct question about the regulatory perimeter itself. The report states that the FCA should review, within the next three to six months, whether large language models such as ChatGPT, Claude and Gemini that provide financial guidance should fall within its regulatory remit. This matters because such tools currently sit largely outside the traditional boundary between regulated financial advice and unregulated guidance, even as growing numbers of consumers use them to compare products, understand financial concepts, and inform major money decisions.
The review frames this as a fundamental regulatory question: if consumers increasingly rely on AI systems provided by a small number of major technology companies to make financial decisions, does it still make sense for those services to sit largely outside the regulatory perimeter that governs traditional financial advice? FCA Chair Ashley Alder described the review as anticipating fundamental change, while stressing that the principles-based, outcomes-focused approach already taken by the FCA on AI, relying on the Consumer Duty and the Senior Managers Regime, has been critical in keeping pace with change so far.
The advice-guidance boundary
UK financial regulation currently distinguishes regulated financial advice, which involves a personalised recommendation and requires FCA authorisation, from guidance, which provides information without recommending a specific course of action. The FCA's targeted support regime already allows firms to offer recommendations to groups of consumers who share similar characteristics, without this counting as full regulated advice. The review notes that AI could extend this further, delivering highly personalised support at scale in a way that could meaningfully close the UK's long-standing advice gap, but that increasingly sophisticated AI may also blur the boundary between advice and guidance in ways that create new consumer protection risks, and open the door to regulatory arbitrage where AI platforms outside the perimeter effectively provide advice-like services without the accountability that comes with authorisation.
Fraud and cyber security risk
Industry reaction to the review has focused heavily on fraud. Commentators have pointed out that most existing fraud controls, including voice callbacks, liveness video checks, and one-off identity document verification, were built to catch a human being pretending to be someone else, not a machine capable of generating a convincing fake of them in real time. The review's warning that fraud and cyber risk will be amplified by AI reflects a genuine capability gap: many firms still treat controls that generative AI can now convincingly defeat as their gold standard, even as synthetic identity fraud grows as a less well-understood threat.
What this means for consumers right now
The Mills Review does not itself create new rules or new consumer rights; it is a review setting out recommendations for the FCA Board to consider, not a policy statement or a piece of legislation. For consumers, the practical takeaway today is less about any immediate change in protection and more about a signal of direction: UK financial regulation is actively grappling with how much trust to place in AI-mediated financial decisions, and whether the tools millions of people already use informally to research financial products should carry the same accountability as a regulated adviser.
Disclaimer: This article summarises an FCA publication and is for general information only. It does not constitute financial or regulatory advice. The Mills Review's recommendations are not yet FCA policy or law; check FCA publications directly for the current regulatory position before relying on any AI tool for a financial decision.
What is the Mills Review?
The Mills Review is an FCA-commissioned review, led by executive director Sheldon Mills, examining how artificial intelligence could transform UK retail financial services by 2030 and beyond. It was published on 6 July 2026.
Will ChatGPT or other AI chatbots become regulated financial advisers?
Not immediately. The review recommends the FCA examine within 3 to 6 months whether large language models offering financial guidance should fall within its regulatory perimeter, but no decision or new rule has yet been made.
How many people already want to use AI for their finances?
An FCA-commissioned survey of over 5,000 UK consumers found 20%, roughly 11 million UK adults, said they would likely use AI capable of acting autonomously within pre-set goals on their personal finances.
What are the main risks the review identifies?
The review highlights four shifts: changes to firm operations, evolving consumer journeys, reshaped market competition concentrated around major technology providers, and amplified fraud and cyber security risk as AI tools become more capable and more widely available.
Does the Mills Review change any rules today?
No. It is a set of recommendations for the FCA Board and Executive to consider, not a new rule, policy statement, or piece of legislation.
Last reviewed: 7 July 2026
