The FX committee asks whether governance can keep up
The body that writes the FX market's rulebook is debating whether its three-year review cycle is still fast enough. Olga Magomedova on what the governance question says about the pace of AI in markets, and about the trader at the desk.
A rulebook asking to be rewritten more often
The Global Foreign Exchange Committee is the international body that coordinates standards for the FX market. It currently updates its best-practice principles for the market once every three years. That cadence was set in a different era. Representatives of the group and other market participants are now debating whether to increase the frequency of reviews for potential technology-related risks.
The venue matters. The push concerns guidelines for the $9.5 trillion-a-day currency market, and the stated reason is to keep pace with artificial intelligence's rapid advance. When the custodians of a market this large start asking whether their own review cycle is too slow, the signal is not about any single product. It is about the gap between the rate at which the tooling is changing and the rate at which the rules around it can be written.
The regulators are asking the same question
The FX committee is not alone. The US Securities and Exchange Commission said on 23 October that the use of artificial intelligence, private assets and tokenised securities would be among key areas of focus for its examination of Wall Street firms in the current fiscal year. The CFTC has moved in parallel. Its forum will examine how advanced AI systems could change derivatives markets, price discovery and risk management, with a session focused on agentic trading, in which autonomous systems conduct financial transactions on behalf of market participants, and the questions raised about accountability, market manipulation and operational failures.
Three separate venues, one shared concern. The question is not whether AI belongs in markets. It is whether anyone can describe, in a stable form, what good behaviour looks like when the tools change every quarter.
What the IMF added to the picture
The macro frame arrived two weeks earlier. The IMF's October 2026 Global Financial Stability Report identified three key risks as rising, each with the potential for broader implications: an artificial intelligence-driven market correction, a sharper rise in sovereign bond yields, and cyberattacks. The AI risk is not a theoretical one about future models. It is a concentration risk about current prices, current flows, and the funding structures around them.
For Magomedova, the useful reading of all three stories together is not about forecasting a correction. It is about understanding that the institutions whose job it is to hold the system steady are telling the public, in plain language, that the environment has changed faster than the controls around it. That is not a reason to panic. It is a reason to prepare.
The gap between the rulebook and the desk
A trader does not operate at the level of the Global Foreign Exchange Committee. A trader operates at the level of the next position. But the two levels are connected. When governance lags technology at the institutional level, the burden of discipline shifts downward. The individual at the desk inherits more of the responsibility for deciding what good behaviour looks like, because the shared rulebook is older than the tools in use.
The engineering reading of this is simple. If the specification lags the hardware, the operator becomes the specification. That is a workable arrangement only when the operator has a written process, honest tolerances, and a habit of reviewing deviations. It fails when the operator assumes the tools have been checked by someone else.
AI is a powerful tool. But tools do not replace responsibility.
What the individual trader can do this quarter
Three habits respond directly to the governance gap. The first is to write down which AI tools are in the trading process, where they sit, and what decision each one is actually making. Many traders cannot answer this when asked. The audit itself is the exercise.
The second is to decide, in advance, which decisions remain the trader's regardless of what any model says. Entry size. Stop placement. Whether to trade at all on a given day. These are the artefacts of the plan, and they do not get delegated to a confidence score produced by a system whose training data the trader has never seen.
The third is to run a short monthly post-mortem on any trade where a model's output and the plan disagreed. Which won. Why. Whether the plan needs an amendment, or whether the model needs to be ignored on that class of setup. The post-mortem is the trader's private version of what the FX committee is debating in public.
The principle that does not move
The headline governance question will be answered in meetings the retail trader will never attend. The desk-level question will not. It will be answered one position at a time, by the person who places the order. The institutional debate about how often to review the rulebook is useful mainly as a reminder that the rulebook on the trader's own desk is the one that gets tested every session.
Build the private rulebook before the public one catches up. Keep the written record. Treat each disagreement between the plan and the model as data, not as a verdict. The governance that matters most to the individual account is the governance the individual trader applies to themselves.