Governing the Rise of Agentic AI in Global Finance

Talal Abu-Ghazaleh

I have recently written about the rise of AI agents and how they are taking over the business sphere in visible and practical ways. We are witnessing the moment where AI is meeting business, and where autonomous systems are moving into useful operational technology. These agents are becoming the machinery of compliance, the engines of liquidity, and the instruments of decision‑making, which demands a supervisory framework equal to the scale of their impact.
The International Monetary Fund (IMF) has already warned that supervisory authorities must themselves adopt AI to oversee AI‑driven markets, identifying explainability, bias mitigation, and governance as essential pillars. The Bank for International Settlements (BIS) has likewise cautioned that central banks face unique risks in adopting AI, including data security, confidentiality, and reputational exposure. The BIS is vital in this conversation as it is the bank, for central banks being the oldest international financial institution, with the ability to convene central banks, harmonize supervisory practices, and anchor global financial stability. By leading the development of governance frameworks for AI adoption, the BIS ensures that innovation does not undermine trust in the global financial system.
The formalization of a framework for agentic AI within the global financial system in particular is therefore a decisive milestone in our civilizational journey toward a secure knowledge economy. Digital agents are becoming central actors in markets, compliance, and risk management, which means that the way we perceive the intersection of technological innovation and systemic risk must change.
Central banks are the anchors of financial stability and their role in this framework is pivotal, as they are now embedding AI into forecasting, payments, supervision, and even banknote production. They are applying adaptive governance frameworks to mitigate risks such as AI hallucinations, bias, and cyber‑attacks. Most importantly, they are coordinating across jurisdictions to align rules and oversight practices. Without harmonization, agentic AI could exploit gaps between national financial regimes, undermining stability. This means that central banks must therefore act not only as regulators but also as global conveners of trust.
The supervisory framework is a necessary condition for innovation to flourish and should not be seen as an obstacle. By embedding agentic AI into regulated structures, we create confidence, enable adoption, and accelerate the transition to a secure knowledge economy. The lessons of history teach us that prosperity depends not only on invention but also on governance. The printing press, the steam engine, and the internet all transformed societies and each required institutions to adapt. Agentic AI is not different and is the next chapter in the story of human progress.
It is my firm conviction that the stability of our future prosperity depends on our ability to govern these AI agents with strategic foresight. This is the beginning of a new era, and if we succeed, we will secure a financial system that is resilient, innovative, and worthy of the trust of generations to come.