FCA Launches Long-Term Review of AI’s Impact on Retail Financial Services
Why the FCA Is Looking Beyond Current AI Rules
The UK Financial Conduct Authority has launched a long-term review into how artificial intelligence could reshape retail financial services, as consumer use of AI tools moves from experimentation into everyday decision-making. The review, led by Sheldon Mills, will report to the FCA Board in the summer and will feed into a public paper intended to frame debate on AI’s role in finance through 2030 and beyond.
Speaking at the FCA’s Supercharged Sandbox Showcase, Mills said the review is focused less on current applications and more on preparing the regulatory system for developments that are still uncertain. “The real challenge in regulation isn’t dealing with what we already understand – it’s preparing for what we don’t,” he said.
The FCA stressed that the work does not introduce new rules or alter its regulatory stance. The authority remains outcomes-based and technology-neutral, with the review designed to test how existing frameworks hold up as AI systems become more capable and more widely used by firms and consumers.
Investor Takeaway
How Consumer Behaviour Is Already Changing
AI has been embedded in financial services for years, particularly in fraud detection, credit decisions, and trading systems. What has changed, according to the FCA, is the pace and visibility of consumer-facing tools. Generative AI, multimodal systems, and emerging AI agents are now being used by households to interpret information and make financial choices.
Mills pointed to survey data showing that AI is already influencing money management. A Lloyd’s survey in 2025 found that one in three customers used AI weekly to . Firms are responding by building tools that personalise guidance, redesign customer journeys, and identify vulnerability earlier in the process.
The regulator’s concern is not that AI adoption is happening, but that neither firms nor regulators yet know which models will dominate or which risks will prove most damaging. “We don’t yet know which models will scale,” Mills said, adding that uncertainty over future risks makes ahead design choices critical.
From Assistive Tools to Autonomous Agents
A central part of the review explores how increasing AI autonomy could change the relationship between consumers and financial firms. The FCA outlined a possible progression from assistive AI, which explains products and highlights risks, to advisory systems that recommend actions such as switching providers or refinancing debt.
The longer-term focus is on autonomous agents that act on behalf of consumers within set boundaries. These systems could move money between accounts, renegotiate contracts, or rebalance savings without direct prompts. Mills described a scenario in which a household AI manages routine financial decisions, reducing administration and avoiding poor-value products.
Such systems, however, raise new questions for accountability and trust. Mills asked how responsibility should be assigned when an AI agent makes a mistake, how consumers can stay meaningfully informed, and how commercial incentives might influence recommendations presented as neutral assistance.
Investor Takeaway
Risks to Consumers and Market Integrity
The FCA highlighted a range of consumer risks linked to more capable AI systems. These include the risk that consumers delegate decisions they do not fully understand, that groups with incomplete data histories face new forms of exclusion, and that fraud becomes harder to detect as criminals adopt AI tools themselves.
Experian data cited by the showed that more than a third of UK businesses reported being targeted by AI-related fraud as ahead as 2024. The regulator warned that fraud techniques are likely to become more convincing and scalable, increasing pressure on firms to invest in defensive systems.
Other concerns relate to bias and explainability. Complex models can produce outcomes that are hard to explain to consumers, while reliance on proxies can lead to uneven results across diverse groups. The data use, transparency, and consent as ongoing pressure points as systems rely on broader and more detailed datasets.
What This Means for Regulation and Competition
The review also examines how AI could alter market structure. AI tools may lower barriers for smaller firms by giving them analytical capabilities once reserved for large banks. At the identical time, access to data and computing power could reinforce the dominance of large incumbents or technology providers that sit outside traditional regulatory boundaries.
These dynamics raise questions for existing accountability regimes. Continuous model updates, rapid scaling of harm, and shared responsibility across challenge assumptions embedded in current supervisory frameworks. The FCA said it will work with other UK authorities, including data protection and competition bodies, to avoid fragmented oversight.
The regulator is asking firms, developers, and market participants to submit views on emerging risks and opportunities by 24 February 2026. Mills closed by urging industry participants to challenge assumptions and highlight blind spots, arguing that regulatory resilience depends on ahead engagement rather than reactive fixes.
