IMRA subcommittee outlines preliminary AI work plan; proposes sector survey and focus on explainability

Gary Kalba, speaking for the Future of Finance subcommittee, presented a preliminary AI work plan focused on two elements: generative AI and machine-learning systems trained with low or no human supervision. Kalba said the subcommittee proposes conducting a survey of AI use in CFTC-regulated markets as an adjunct to the Commissions existing request for comment, and may formulate recommendations ranging from guidance and advisories to potential rulemaking.

Key areas the plan would explore include disclosure or explainability requirements for registrants using AI; testing and monitoring regimes (cybersecurity, data controls, bias mitigation, margin and model output consistency); and governance frameworks that could include designated personnel or senior-management oversight responsibilities. Kalba emphasized the subcommittee will seek diverse expert input, coordinate with other regulators (SEC, Federal Reserve) and leverage responses to the CFTC's RFI.

Members raised scope and burden concerns during discussion: some urged careful scoping to avoid capturing routine algorithmic tools already covered by existing principles-based frameworks, while others endorsed an early, principles-focused approach to identify whether AI presents qualitatively different risks. The subcommittee will continue to refine the plan, consult experts, consider a voluntary versus mandatory survey, and report back to IMRA.

What happens next: the Future of Finance subcommittee will develop the survey design and continue coordinated outreach with regulators and subject-matter experts; the subcommittee intends to use survey and RFI responses to refine any recommendations for the Commission.