The American Medical Association released a governance framework on October 15, 2026, establishing accountability guidelines for agentic AI assuming clinical functions. The release comes alongside international actions by the Financial Stability Board to address regulatory gaps left by US banking regulators, who explicitly excluded autonomous AI models from model risk guidance. Industry reports highlight widespread unauthorized deployments and inadequate kill-switch preparedness.
AMA framework for agentic healthcare AI
- ▪The American Medical Association released a new governance framework on October 15, 2026, to guide the safe deployment of agentic AI assuming clinical functions in healthcare.
- ▪The American Medical Association's framework focuses on four questions covering clinical functions, evidence of performance, conditions of permission, and risk accountability.
Survey of US healthcare leaders on AI
- ▪In a survey of 250 U.S. healthcare leaders, 28% reported deploying agentic AI tools and 72% reported that agentic AI is occasionally implemented without IT approval.
- ▪In a survey of 250 U.S. healthcare leaders, 44% of respondents reported conducting pilots or proof-of-concept projects with agentic AI tools.
Financial Stability Board guidance on AI adoption
- ▪Sound Practice 10 of the Financial Stability Board's guidance establishes a risk-based intervention model requiring mandatory human sign-off for financial transactions exceeding specific thresholds.
- ▪The Financial Stability Board declined to framing AI systems as synthetic employees, citing concerns that the analogy obscures unpredictable autonomous decisions.
- ▪The Financial Stability Board published 12 non-binding recommended practices on October 16, 2026, for the responsible adoption of artificial intelligence in financial institutions across G20 economies.
FSB Sound Practice 12 requirements
- ▪Sound Practice 12 expects financial institutions to secure contractual rights covering audit access, data portability, security incident notification, and orderly service termination before deploying externally sourced AI systems.
- ▪Updated Sound Practice 12 requires financial institutions to perform due diligence and maintain ongoing oversight of embedded AI functionality extending to fourth- and nth-party vendors.
US model risk guidance
- ▪US regulatory guidance OCC Bulletin 2026-13, issued jointly by the OCC, Federal Reserve Board, and FDIC on April 17, 2026, explicitly excluded generative and agentic AI models from its scope.
- ▪The OCC, Federal Reserve Board, and FDIC plan a future request for information regarding bank AI model risk management, though no timeline has been set.
Andrew Bailey on frontier AI risks
- ▪Bank of England Governor Andrew Bailey described frontier AI threat capabilities as the most immediate cyber risk facing global financial systems in an August 31, 2026 letter.
- ▪Bank of England Governor Andrew Bailey proposed pre- and post-deployment testing protocols for frontier AI models used in finance and G20-level agreements governing access to foreign-built frontier models.
Moody's warning on AI vendor dependency
- ▪Moody's warned in August 2026 that banks deploying AI risk systemic dependency on concentrated technology providers, specifically naming OpenAI and Anthropic.
- ▪Moody's warned in August 2026 that technology providers OpenAI and Anthropic face investor pressure to reach profitability, creating potential pricing leverage over dependent financial institutions.
Story comments
Loading comments…