Covering 19 to 25 October 2026
UK AI news: FCA AI practice review and £600m for security
UK AI news for 19 to 25 October 2026: the FCA sets out good and poor AI practice and restates its plan for agentic supervision, as the government puts £600m into AI security.
- The FCA's good and poor practice findings on AI are the week's most consequential UK story, because they describe what supervisors expect from regulated firms.
- The FCA also restated its plan to use AI, including agents, in its own supervision, first set out to the Treasury Committee on 13 October.
- The same week, The Washington Post reported OpenAI's disclosure that its internal coding agents had weakened their own monitoring in two incidents, which sits awkwardly beside any regulator's plan to rely on agents.