What happened
On 19 October 2026 the FCA published a new webpage (with an accompanying speech by chief data, information and intelligence officer Jessica Rusu) setting out good practice and areas for improvement in firms' use of AI in financial services, drawn from industry engagement, supervisory work and its AI Lab (AI Live Testing, Supercharged Sandbox). It covers consumer duty and vulnerable customers, governance and accountability (incl. senior management functions, human oversight), and operational/cyber resilience and third-party relationships, explicitly noting it does not create new requirements or prescribe a single model. It also calls on AI providers to add clear disclaimers about the limits of model-generated financial information/advice.
Why it matters
This is the FCA's first consolidated good/poor-practice guidance on AI (including generative and agentic AI) in financial services, giving UK regulated firms an authoritative benchmark for proportionate AI governance, human oversight and cyber/operational resilience. Though not a new rule, it signals supervisory expectations and can be used to anticipate FCA AI supervision.
Action needed
UK financial-services firms using AI should review the good/poor-practice examples against their AI governance, oversight and operational-resilience controls, and align AI-generated consumer-facing content with the FCA's transparency expectation.