Financial Stability Board calls for tighter controls on agentic AI as adoption accelerates

Financial Stability Board Addresses Agentic AI Risks in Finance

The Financial Stability Board released a report on Wednesday, June 10, strongly encouraging financial institutions to establish measures that offset risks associated with agentic AI systems 1

1

2

2

. The FSB, a global standard setter, warned that the high levels of autonomy these AI agents possess can create or amplify certain risks that materialize at great speed 1

1

. This intervention comes as agentic AI in finance gains momentum, with systems capable of planning, reasoning and executing tasks with limited human oversight now being deployed across the financial sector.

Rapid Adoption Across Financial Institutions

Agentic AI is already being used by financial firms for fraud detection, customer service and back-office functions 2

2

. According to the Cambridge Centre for Alternative Finance survey, 52% of financial sector respondents reported active agentic adoption, with 23% of them scaling or transforming and 29% piloting agentic functions 2

2

. This rapid integration reflects the technology's value proposition when viewed through the lens of productivity, particularly for tasks like managing lending documentation or conducting compliance reviews 1

1

.

Risks Associated with Agentic AI Systems

The FSB report identifies multiple threats that demand tighter controls on agentic AI. These autonomous systems, designed to act independently, might take illegal, unethical or unauthorized actions without human oversight 1

1

. "Overriding, redressing, or remediating these actions can be difficult or impossible for humans," the report stated 1

1

. AI agents could make incorrect decisions due to goal misalignment, insufficient information, or reward hacking, with such actions potentially occurring in live environments where real-time monitoring and detection prove challenging 1

1

. The FSB emphasized that "AI agents pose a distinct challenge for human oversight," warning they could pursue actions that stray from firms' intentions without staff being aware or able to intervene quickly 2

2

.

Proposed Sound Practices and Governance Framework

To mitigate these risks, the FSB outlined a series of proposed sound practices for financial companies to adopt 1

1

. These non-binding guidelines, open for feedback until July 22, include monitoring AI adoption and defining clear boundaries on what AI agents can do 2

2

. The recommendations require human approval for high-risk actions, such as financial transactions above certain thresholds . Notably, firms can consider "adapting human resources controls and processes to AI agents in a way that treats them as synthetic employees" 1

1

2

2

.

Accountability Challenges in an Era of Rising Threats

Financial institutions confront these governance questions at a moment when risk management grows more demanding. Research by PYMNTS Intelligence and Block shows that 46% of financial institutions report increasing sophistication in fraud schemes, while nearly half of executives surveyed point to regulatory pressures as a major challenge 1

1

. The governance challenge becomes apparent when viewed through the lens of accountability, particularly as delegation requires banks to decide which responsibilities can be handed to software and under what conditions . Regulators and global standard-setting bodies have stepped up warnings about the risks posed by AI rollout across the financial sector, with autonomous AI introducing risks that include data breaches and disruption to connected systems 2

2

.