FCA launches forward-looking review into AI's transformative role in insurance and finance

The Financial Conduct Authority has opened a forward-looking review into how artificial intelligence could reshape insurance markets and retail financial services over the coming years, signalling heightened regulatory attention as firms deepen their use of machine learning across pricing, underwriting and claims handling.

According to the FCA, insurers are already deploying AI for continuous risk assessment and automated claims processing in ways that are materially affecting operations. The authority said it expects further, more profound changes as the technology matures and becomes cheaper and more widely adopted. The review, led by Sheldon Mills, will examine likely technological trajectories, the implications for firms and consumers, and whether current regulatory frameworks remain fit for purpose. Findings will inform recommendations to the FCA board at its summer meeting, and industry stakeholders have until 24 February to submit views.

The scope excludes wholesale markets and largely stops short of wider societal impacts, although the FCA said it will consider indirect social effects where they have a bearing on retail financial outcomes. The exercise is not intended to produce new, prescriptive rules at this stage; rather, the regulator will test whether existing rules and supervisory approaches can accommodate the changes AI is likely to bring. The FCA has already signalled an interest in how AI might alter supervision itself, including supervisory techniques and tools.

The review follows a series of public warnings from senior regulators about AI risks. In September 2024, FCA chief executive Nikhil Rathi cautioned that AI-enabled “hyper-personalisation” in insurance could leave some consumers effectively uninsurable and exacerbate discriminatory outcomes, particularly for people with certain health conditions or limited digital access. According to reporting by The Guardian in January 2026, MPs on the Treasury Committee have urged the FCA, the Bank of England and the government to move beyond a largely “wait-and-see” stance and adopt a more proactive approach to mitigating AI-related harms in finance.

The FCA’s concern about both consumer protection and market functioning is reflected in specific areas under scrutiny. AI-driven agents that compare policies and assist customers in switching providers could reduce customer inertia and benefit some consumers, but they also threaten business models that rely on sticky customers and cross-subsidisation of risk pools. At the same time, the authority has warned that advances in AI may enable more sophisticated fraud and market manipulation, creating a fresh set of operational and conduct risks for firms and supervisors to address.

This initiative builds on the FCA’s earlier research into firm practices. A 2023 FCA note found that about three quarters of UK financial firms were using AI and that most had assigned individual accountability for AI approaches; more than half of reported use cases involved some form of automated decision-making, with cybersecurity cited as the leading risk. Separately, the FCA announced in December 2025 the AI Live Testing programme to help firms trial AI applications safely in collaboration with the regulator, underscoring an appetite for controlled, evidence-based experimentation.

The Prudential Regulation Authority, running a parallel but distinct supervisory agenda, has published priorities for 2026 that include oversight of high-demand pension risk transfers and a semi-live crisis exercise, underlining that safety and resilience concerns remain prominent across UK financial regulators.

Industry groups and consumer advocates are likely to press the FCA on trade-offs between innovation and fairness. The regulator faces difficult judgments: interventions that preserve broad access and prevent discriminatory outcomes could constrain some AI-enabled efficiencies, while a lighter touch risks entrenching harmful segmentation of cover and pricing. The FCA’s choice to focus on forward testing rather than immediate rulemaking suggests it seeks to strike a balance, using evidence from pilots, firm submissions and industry testing to shape any future supervisory or rule changes.

As the review proceeds, firms will need to demonstrate how they identify and mitigate model risks, ensure explainability and fairness in automated decisions, and protect customers from novel forms of harm. The regulator’s workstream is poised to influence industry practices through guidance, supervisory expectations and, ultimately, potential rule changes once the FCA assesses whether the existing regime can manage the systemic and consumer risks AI may introduce.