UK Financial Services Knows Where AI Risk Sits. The Controls Haven’t Caught Up.
Blog post from Duality
A commentary on the Bank of England and FCA’s 2026 survey of 196 financial-services firms finds AI adoption approaching universality, with 86% of respondents using it and all surveyed banks reporting use, while data protection, security, and prompt or output leakage remain leading concerns. Third-party AI is becoming more significant, representing 39% of reported systems and 53% of high-materiality systems, yet firms often lack transparency into vendors’ operations and rely heavily on contractual assurances rather than direct technical evidence. The author argues that confidential computing, using hardware-protected environments and cryptographic attestation, could provide verifiable safeguards for data, prompts, retrieved context, and models during processing, although it would not address model hallucinations, bias, explainability, output-access policies, or firms’ accountability. The piece recommends that organizations inventory high-materiality third-party AI systems, assess what data leaves their environment and who can access it, and seek evidence that handling practices can be verified and enforced rather than merely promised.
No tracked trend matches for this post yet.
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.