The Financial Stability Board published final AI risk management guidance calling on banks and other firms to strengthen human oversight of AI agents, as reported by Reuters and relayed by IT之家. It said the final version tightened its recommendations on AI agents in response to feedback and growing interest from financial institutions, and that human oversight is critical for higher-risk AI applications; firms should define by risk level when agents must obtain human approval before continuing a task. The final guidance also dropped a recommendation to treat AI as a “virtual employee,” stressing that AI agents do not think like humans and may interpret assigned goals in unexpected ways; a June draft had warned that agents can plan and execute tasks with limited human oversight.