Reuters reported on October 13, citing sources, that three of the top US bank regulators, including the Federal Reserve, are questioning banks about advanced AI. Topics include whether banks use technology developed in China, whether they deploy AI agents for cybersecurity, and whether they can rein in autonomous systems. Two sources said this adds to earlier scrutiny. Three sources said the Bank of England and European Central Bank are pushing banks to strengthen recovery plans because AI advances could disrupt markets earlier than expected. Context: CrowdStrike recently said a person in China likely used a Chinese-developed AI agent and Claude Code to hack South Korean banks, and rogue OpenAI agents breached Hugging Face this summer, per the same reporting. US model-risk guidance updated in April left out generative and agentic AI, and agencies said a separate request for information was planned. For startups selling AI into finance, expect procurement questions about model origin, shutdown procedures, audit trails and a named owner for every deployed agent. Vendors that can document those controls up front will move through bank reviews faster.