The UK's Financial Conduct Authority is calling on financial services firms to report AI failures and near misses, according to The Register. The regulator's stated aim is clear accountability as the industry increasingly relies on autonomous agents to make or assist decisions.
This regulatory push reflects the growing role of compute and AI infrastructure in finance, where system errors can have direct consequences for consumers and markets. By requiring disclosure of incidents and close calls, the FCA appears to want firms to demonstrate that they understand and can govern the AI systems they run.
The source does not detail specific thresholds, reporting formats, or technical standards, so the practical impact on hardware and software pipelines remains unclear. Still, the direction is notable: financial firms may need to build more robust monitoring and audit trails into their AI compute stacks to satisfy the watchdog's expectations.