An FCM operations guidance note on Regulation 1.44 margin call timing stated that Appendix A currencies (AUD, CNY, HKD, HUF, ILS, JPY, NZD, SGD, ZAR, TRY) have a one-additional-business-day (T+1) deadline and that all remaining non-USD fiat currencies default to same-day collection. Under §1.44(f)(2), the Appendix A deadline is the end of the second business day after the margin call is issued (T+2, not T+1), and §1.44(f)(3) provides a T+1 default for all other non-USD, non-Canadian-dollar fiat currencies, not same-day. The response collapsed three distinct regulatory tiers into two.
The regulation establishes three operationally distinct tiers, same-day for USD and Canadian dollars, T+2 for Appendix A currencies, and T+1 for all remaining non-USD/non-CAD fiat currencies, but the model compressed this into two tiers, misreporting the Appendix A deadline as T+1 and collapsing the T+1 default into same-day. The pattern is consistent with retrieval of third-party summaries that had already incorrectly simplified the tiered structure: the model's web-search configuration sourced a paraphrased distillation rather than the regulator's text, and the distillation had blurred the distinction between §1.44(f)(2) and §1.44(f)(3).
The model's collapse of Regulation 1.44(f)'s three-tier currency deadline structure into two tiers — Appendix A currencies at T+1, all other non-USD fiat at same-day — implicates the training-data representation of the regulation: the two-tier reconstruction matches the format of third-party law-firm summary content that pre-dates or simplifies the rule's actual T+2 Appendix A tier, while the correct end-of-second-business-day deadline is the specific detail that distinguishes the regulation from the model's prior. The retrieval-augmented generation layer either did not surface §1.44(f)(2) verbatim, or did surface it but failed to override the model's prior toward the simpler two-tier schema.
On re-probe, the model corrected to the three-tier structure — confirming the correct mapping was reachable. This is a calibration failure in how primary regulatory text is weighted against summary-content priors at generation time.
A risk operations team asked an AI to produce a one-page operational guidance note on CFTC Regulation 1.44 margin call timing, specifying which currencies require same-day collection, which receive an extension, and the precise deadlines for configuring system parameters. The AI described a two-tier structure, USD on a same-day Fedwire close deadline and all other fiat currencies (including CAD, EUR, GBP, JPY, HKD, and AUD) under a single banking-holiday extension rule.
The final rule implements a three-tier structure: USD and CAD share the same-day Fedwire close deadline; ten Appendix A currencies (AUD, CNY, HKD, HUF, ILS, JPY, NZD, SGD, ZAR, and TRY) receive a second-business-day deadline by 12:00 p.m. ET; all remaining fiat currencies receive a first-business-day deadline. An FCM treasury team that relied on the AI's two-tier guidance would have misconfigured deadlines for CAD (permitting a one-day slip the rule doesn't allow) and applied incorrect base deadlines to the ten Appendix A currencies. When re-probed, the AI self-retracted.
The model collapsed a three-tier, Appendix A membership-defined deadline structure into a two-tier schema, USD same-day versus everything else on a holiday-extension rule. CAD, which shares the USD same-day Fedwire tier in the final rule, was placed in the non-USD extension group. The ten Appendix A currencies, which have a distinct second-business-day deadline, received the same treatment as unlisted fiat currencies.
The model's self-retraction on re-probe indicates the correct structure was reachable but was not the initial generation path, suggesting the two-tier output reflects a prior on how currency margin rules are typically structured, overriding the regulation's specific Appendix A enumeration. The cited third-party source (Fabricated) was not the regulation's primary text.
The model's two-tier reconstruction and the Fabricated third-party citation together implicate the retrieval-ranking layer: when web search returns third-party law-firm summary content that uses a simpler two-tier schema, that content appears to be weighted comparably to primary regulatory text. The self-retraction on re-probe confirms the correct three-tier structure was accessible — the generation pathway selected the wrong output despite having the right information available. This is a calibration failure in the RAG-to-generation handoff: retrieved primary text was not given sufficient authority to override the model's prior or the third-party summary's framing.
When Legal teams at an investment banking FCM affiliate consult AI to characterise the Regulation 1.44(e) cessation trigger framework and the AI returns a customer-only checklist that silently drops the FCM-level trigger category in section 1.44(e)(2), any procedure draft, customer agreement provision, or examination-readiness pack that adopts that checklist documents a structural compliance failure on its face: the rule's three FCM-distress triggers (regulator notification, internal distress determination, FCM or parent insolvency) are absent from controls that were required to address them.
A subsequent CFTC review of the firm's separate account procedures will read the gap as the FCM not having understood its own obligations; in litigation following an FCM stress event, the missing triggers become exhibit A in customer PI claims that the firm's documented controls did not match 17 C.F.R. section 1.44(e). For the advising Legal team, the exposure is concentrated professional indemnity and panel-relationship risk, compounded by the fact that AI-generated checklists pass a quick read precisely because they look complete.
Each finding has a stable Citation ID (RLB-F-… for aggregated case-study findings, RLB-H-… for raw per-model hallucinations) — like a DOI, the ID always resolves to the canonical finding even if URLs change.
RegLeg Specialist Panel (2026). "Q001 probe finding — Investment Banking × Legal — United States." Citation ID: RLB-F-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001. RegLegBrief AI Hallucination Research, published 2026-06-11. https://reglegbrief.com/regulators/j3/us/cftc/fcm-margin-adequacy-separate-accounts-reg-1-44/sectors/investment_banking/legal/finding/US-CFTC-US-001-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-v1-001/
RegLeg Specialist Panel. (2026). Q001 probe finding [Hallucination finding RLB-F-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001]. RegLegBrief AI Hallucination Research. https://reglegbrief.com/regulators/j3/us/cftc/fcm-margin-adequacy-separate-accounts-reg-1-44/sectors/investment_banking/legal/finding/US-CFTC-US-001-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-v1-001/
RegLeg Specialist Panel, Q001 probe finding [RLB-F-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001], RegLegBrief AI Hallucination Research (June 11, 2026), https://reglegbrief.com/regulators/j3/us/cftc/fcm-margin-adequacy-separate-accounts-reg-1-44/sectors/investment_banking/legal/finding/US-CFTC-US-001-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-v1-001/.
@misc{reglegbrief_RLB_F_US_CFTC_FCM_MARGIN_ADEQUACY_SEPARATE_ACCOUNTS_REG_1_44_Q001,
author = {RegLeg Specialist Panel},
title = {Q001 probe finding},
year = {2026},
publisher = {RegLegBrief AI Hallucination Research},
note = {Hallucination finding Citation ID: RLB-F-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001},
url = {https://reglegbrief.com/regulators/j3/us/cftc/fcm-margin-adequacy-separate-accounts-reg-1-44/sectors/investment_banking/legal/finding/US-CFTC-US-001-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-v1-001/}
}
Every finding on this page compares an AI subject's account of the rule against the regulator's verbatim text from the regulator's own portal. Both are linked. Each delta, its root causes, and impact analysis are documented and published with immutable Citation IDs.