FCA Releases Guidance on AI Practices in Finance
Start here
AI-assisted source synthesis · not an independent investigation.
On October 19, 2026, the UK Financial Conduct Authority (FCA) said it would work with the financial industry, other regulators, and frontier laboratories to jointly develop a financial regulatory framework for AI agents, covering issues such as agent identity, delegated authority, responsibility, and enforcement. The FCA also released good and poor AI practices, requiring financial institutions to put consumers first, establish clear accountability and human oversight, and prepare for system failures, supplier dependence, and major failures or near misses. On the regulatory front, the FCA is using AI for market supervision, including developing tools for approximately 60,000 anti-money-laundering entities planned to come under its regulatory scope in 2028, and supervising approximately 9,000 small wholesale-market transaction-reporting entities through agent tools.
AI-written from reports · updated 1 hour ago
Timeline
Follow the reporting to see the different sides of the story.
- FCA (UK)SelectedFCA将与业界共同开发智能体金融监管框架
FCA said it will work with the financial sector, other regulators and frontier labs to jointly develop an agentic finance framework covering agent identity, delegated authority, accountability and execution. FCA published good and poor practices for AI, requiring firms to put consumers first, establish clear accountability and human oversight, and prepare for system failures, vendor dependencies and major failures or near misses. FCA is also using AI to supervise markets, including developing tools for the 60,000 AML firms that will come within its regulatory perimeter in 2028, and using agentic tools to oversee approximately 9,000 small wholesale market transaction reporting firms.
Heat over time
Not enough consecutive observations yet to draw a trend.