10 February, 2026 (London/Singapore): The FIX Trading Community, the industry association that manages the world’s trading language, the FIX Protocol, has proposed ten key recommendations for regulating AI in trading, in response to the Monetary Authority of Singapore’s (MAS) consultation, Consultation Paper on Guidelines on Artificial Intelligence Risk Management, which closed on January 31.
Executive Director Jim Kaye said the consultation was both valuable and timely, enabling FIX and its member firms to come together to address the complex issue of AI, with which all firms around the world are currently grappling.
“This was an immensely useful exercise in teasing out both overarching themes and technical challenges,” he said. “It’s allowed us to propose some practical measures that will be very helpful to firms as we, as an industry, meet the risks and opportunities that AI presents.”
Machine learning has been the backbone of algorithmic trading for decades, but the advent of large language models, and high-profile examples of their failings, have thrown AI into the spotlight. The interconnected nature of global markets means that one AI catastrophe could potentially have far-reaching consequences, and the speed at which the AI industry is moving means both regulators and participants are highly focused on this topic.
“This consultation allowed us to think about AI from all angles – from Board accountability (critical) to existing frameworks that could be used to help manage AI risks,” Kaye said. “The MAS proposals are an excellent start – we think that with some additional depth, they will be a good framework for firms to manage risk as AI adoption moves ahead.”
FIX’s recommendations include ten key points:
- Agreeing a globally recognised definition and taxonomy for what “AI” means in a trading context. Without this, and particularly if firms are left to define AI for themselves, there will be inconsistency between both firms and markets, and regulatory arbitrage. Standards and publications that can be referenced include Coalition for Content Provenance and Authenticity (C2PA) and ISO 24030.
- Making Boards accountable for AI governance effectiveness – not just AI governance existence.
- Anchoring AI regulatory guidelines to existing constructs for operational resilience and market integrity, such as MiFID RTS 6 which requires firms to both annually self-assess and validate their algo trading systems as well as to certify non-contribution to market disorder on every significant change; and the EU’s Digital Operational Resilience Act (DORA) ICT risk management frameworks.
- Specifically considering the way large language models (LLMs) are integrated into algorithms and calculators, including potential downstream workflow and decision-making impacts.
- Establishing cross-functional committees to ensure risk management across the organisation, and ensuring these committees also have oversight of third-party AI-related suppliers.
- Creating internal and regulator-facing disclosure patterns analogous to food labelling: what data was used, training approach, evaluation results, mitigations, limitations, intended use, and user-data protection.
- Supporting the creation and maintenance of AI inventories, the granularity of which will necessarily be well beyond typical algo inventories.
- Imposing the proposed risk dimensions of impact, complexity and reliance with the addition of a further two dimension: interconnectedness/contagion potential; and change sensitivity/nonlinearity – ie the butterfly effect, where small changes can lead to material differences in AI behaviour.
- Strengthening the proposed standards, processes and controls through the AI lifecycle. FIX Trading Community’s significant work on algorithm testing and certification could be referenced here.
- Revising the definition of “significant/material change” so that retraining, data refreshes, fine-tuning, prompt/policy-layer changes, and infrastructure/latency changes are each evaluated under a clear, risk-based materiality test, that may require re-approval and re-testing.
“There is no doubt that Silicon Valley’s ‘move fast and break things’ approach is at odds with the conservative approach that must be taken to protect the integrity of our global capital markets,” Kaye said. “When it comes to AI, it’s critical that the entire capital markets industry works together to ensure we protect our financial system as well as we possibly can.”
The FIX Trading Community is the only independent global community where capital markets firms come together to solve common issues and shape the evolution of capital markets. FIX groups in over 60 countries are working on a range of global issues including digital assets; reference data; carbon trading; AI; algo trading; FICC and ETFs, while country and regional committees work together to manage local regulation and market structure matters. To see what FIX can do for your firm, visit www.fixtrading.org.
For further information and interviews with Jim Kaye:
Kristin Westlake, The Continuum Partners | kwestlake@thecontinuumpartners.com | +61 416 219 358/+1 (604) 966 1410