Central banks and foreign exchange participants are moving to review the governance of the currency market more frequently, so their rules don’t fall behind artificial intelligence.
That matters because the FX market is estimated to handle more than $5 trillion in daily turnover.
Why regulators are rethinking the rulebook
The concern centers on agentic AI. These are autonomous systems that don’t just suggest trades but can make independent trading decisions on their own.
The Bank of England has flagged the potential for herding, where many AI agents pile into the same trade at once. It has also warned about volatility spikes and possible conflicts with established monetary policy goals.
Bank of England Deputy Governor Sarah Breeden addressed agentic AI’s implications in June 2026. She pointed to the need for market-wide guardrails, rather than leaving every firm to police its own machines.
The UK’s Financial Conduct Authority and the Bank of England are both assessing two specific tools. One is runtime governance, meaning oversight that operates while an AI agent is actually live and trading. The other is kill switches.
The FSB sets the global baseline
At the international level, the Financial Stability Board published its final report on October 15, 2026. The document is titled “Sound Practices for Responsible Adoption of Artificial Intelligence.”
The report was directed at G20 Finance Ministers and Central Bank Governors. It lays out 12 key practices for deploying AI responsibly.
Three themes run through it. Governance should be organization-wide, not siloed in a tech team. Risk management should be rigorous across the entire AI lifecycle, from design to retirement. And high-risk AI applications need robust human oversight.
Simulation work by the BIS Innovation Hub and other regulatory bodies suggests that market-wide safeguards, such as circuit breakers, might need to be introduced to contain the risks autonomous systems bring.
How banks are responding
UBS has implemented a risk-tiered governance structure for its AI and machine learning work in FX trading. A model that tidies up internal reports doesn’t need the same scrutiny as one that executes client trades. UBS has aligned this approach with the FX Global Code.
UBS received the Euromoney award for best use of AI/ML in FX in 2026, with its governance framework highlighted as part of that win.
The scale problem
The commonly cited figure puts daily FX turnover above $5 trillion, while more recent estimates cited for the 2025-2026 period range between $7.5 trillion and $9.6 trillion.
What this means for the market
There is a real tension built into all of this. Market-wide guardrails like circuit breakers could reduce the risk of AI-driven herding and sudden spikes. They could also interrupt liquidity at exactly the moment traders need it most.
For traders and investors, the key items to watch are concrete. Will the FCA and Bank of England turn their assessment of kill switches and runtime governance into binding requirements? Will any FX-specific circuit breaker proposal emerge from the BIS simulation work? And how quickly will the FX Global Code itself be updated to reflect agentic AI?
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.