The Financial Stability Board released a report on Wednesday, June 10, calling for tighter controls on agentic AI systems in the financial sector. The FSB cited risks from autonomous AI that can materialize at great speed, including unauthorized actions, data breaches, and disruption to connected systems. The guidance responds to accelerating adoption of AI capable of planning, reasoning, and executing tasks with limited human oversight across fraud detection, customer service, and back-office functions.
Financial Sector Reports 52% Active Agentic AI Adoption
According to a Cambridge Centre for Alternative Finance survey, 52% of financial sector respondents reported active agentic AI adoption. Of these, 23% are scaling or transforming operations, while 29% are piloting agentic functions. Agentic AI refers to systems capable of planning, reasoning, and executing tasks with limited human oversight.
Regulators and global standard-setting bodies have increased warnings about AI risks since Anthropic released Mythos, which experts view as posing significant cybersecurity challenges to the banking industry.
FSB Identifies Speed and Autonomy Risks in AI Systems
The FSB report stated that autonomous AI introduces risks that can "materialize at great speed." The standard setter warned that AI agents could pursue actions that stray from firms' intentions without staff being aware or able to intervene quickly. The report noted that "AI agents pose a distinct challenge for human oversight."
Identified risks include the possibility of unauthorized or illegal actions, data breaches, and disruption to connected systems.
Proposed Safeguards Include Human Approval Thresholds and Feedback Deadline
The FSB outlined a series of proposed "sound practices" in non-binding guidelines open for feedback till July 22. The recommendations urge financial firms to define clear boundaries on AI use and embed safeguards.
Proposed measures include boundaries on what AI agents can do and requirements for human approval for high-risk actions, such as financial transactions above certain thresholds. The FSB also suggested firms consider adapting HR controls and processes to AI agents in a way that treats them as "synthetic employees."
FAQ
What did the Financial Stability Board release on June 10?
The Financial Stability Board released a report on Wednesday, June 10, calling for tighter controls on agentic AI systems in the financial sector. The report outlined proposed "sound practices" including boundaries on AI agent actions and human approval requirements for high-risk tasks.
How widely is agentic AI adopted in the financial sector?
According to a Cambridge Centre for Alternative Finance survey, 52% of financial sector respondents reported active agentic AI adoption. Of these, 23% are scaling or transforming operations, while 29% are piloting agentic functions. Agentic AI is used for fraud detection, customer service, and back-office functions.
