FSB urges financial institutions to impose strict controls on AI deployment

The Financial Stability Board (FSB) has called for tighter controls on AI deployment in financial services, warning that new technologies such as agentic AI bring unique risks.

In its report Sound Practices for Responsible Adoption of Artificial Intelligence (AI), the global financial watchdog lays out 12 “sound practices” for responsible AI adoption in financial services.

These recommendations cover governance, lifecycle management, and risk management requirements that the watchdog sees as essential to safe AI deployment.

“The board and senior management of financial institutions are strongly encouraged to reference the sound practices as they consider business strategy, technology adoption, and risk management in an increasingly AI enabled environment,” the regulator wrote.

On governance, the FSB urged financial leaders to align their AI adoption with their organisation’s strategy and risk appetite, adopt risk management frameworks that address potential AI risks, and continue to adjust these measures as AI evolves.

The watchdog also toldfinancial institutions to carefully assess whether to adopt third-party AI models or develop their own in-house, where “traditional” AI can be used rather than agentic AI, and how explainable their AI models are.

All of this, the FSB added, must be underpinned with proper data governance and human oversight of AI activities.

It also cautioned that AI models present their own risks, with sound practices 11 and 12 recommending financial institutions adapt their cyber risk plans to assess the dangers of AI deployment and properly manage the risks associated with third-party AI use.

On the risks that AI agents pose in particular, the FSB wrote: “An AI use case can pose high risks even if it is not material. For example, an AI agent performing a low materiality task may still have access to ICT assets or a significant amount of sensitive data.

“Consequently, the agent could cause significant harm to the financial institution or its customers if it accidentally or intentionally acts outside defined parameters.”

The FSB acknowledged that AI has rapidly evolved since its 2024 report on how the technology will impact the financial services sector and noted that it may have to reassess its recommendations in the near future.



Share Story:

Recent Stories


Creating value together: Strategic partnerships in the age of GCCs
As Global Capability Centres reshape the financial services landscape, one question stands out: how do leading banks balance in-house innovation with strategic partnerships to drive real transformation?

Data trust in the AI era: Building customer confidence through responsible banking
In the second episode of FStech’s three-part video podcast series sponsored by HCLTech, Sudip Lahiri, Executive Vice President & Head of Financial Services for Europe & UKI at HCLTech examines the critical relationship between data trust, transparency, and responsible AI implementation in financial services.

Banking's GenAI evolution: Beyond the hype, building the future
In the first episode of a three-part video podcast series sponsored by HCLTech, Sudip Lahiri, Executive Vice President & Head of Financial Services for Europe & UKI at HCLTech explores how financial institutions can navigate the transformative potential of Generative AI while building lasting foundations for innovation.

Beyond compliance: Building unshakeable operational resilience in financial services
In today's rapidly evolving financial landscape, operational resilience has become a critical focus for institutions worldwide. As regulatory requirements grow more complex and cyber threats, particularly ransomware, become increasingly sophisticated, financial services providers must adapt and strengthen their defences. The intersection of compliance, technology, and security presents both challenges and opportunities.