Technology Transactions Soundbite | Inside the Bank of England/FCA 2026 survey on AI in financial services
Tech goes mainstream, but risks cause concern
The fourth survey of AI and machine learning in UK financial services shows that AI deployment is now routine among the 196 participating firms and is widespread in some. Use of genAI and higher materiality models is gradually increasing, although the autonomy of agentic systems remains tightly controlled, particularly for client-facing purposes. In terms of risk, the survey showed:
- More firms are deploying third party systems, but the market is highly concentrated. The three largest providers dominate cloud, compute, and API/hosting services, resulting in critical third-party dependency and operational risks. Negotiating adequate contractual protections in areas such as liability, audit rights and data usage can prove difficult.
- Firms are also grappling with a lack of transparency of third-party systems, particularly for training data, model design, performance limitations, and failure modes.
- All risk drivers – data, cyber, supply chain, operational resilience and genAI-specific issues – are expected to amplify in the next 1-3 years. Despite having governance and risk management measures in place, firms are finding effective oversight and monitoring to be an ongoing challenge.
All respondents recognised the transformative benefits AI is bringing to the sector in efficiency and productivity, and more are anticipated - including enhanced cybersecurity, operational resilience and competitive positioning. However, clear guidance on best practice is seen as crucial to remove constraints to adoption and the sector echoes the calls in Mills for agile and future-proof frameworks so that AI's full benefits can be realised at speed and scale.
Authors: Fiona Ghosh, Partner; Patricia Wade, Expertise Counsel
