Enumeration Collapse on a CFTC margin rule: two frontier AI models rewrote the currency deadline tiers on Regulation 1.44, and a law firm deliverable built on either output would direct the client to a standard the regulation does not contain.

The exposure runs through 2-page client memos, partner briefings, and the firm's internal knowledge-base entries on Regulation 1.44 that are now seeded with the wrong tier structure unless human review has reset them.

The pattern in one line

Frontier AI models, asked to map CFTC Regulation 1.44's currency deadline tiers for a usable client deliverable, produced two structurally wrong deadline frameworks: one collapsed three tiers into two, the other added an intraday cutoff the rule does not contain. A law firm memo, briefing, or training note built from either output would document a standard the CFTC has not set and would create a written firm position that opposing counsel or an examiner could cite back against the client.

How the RLB Specialist Panel tested this

Questions are prepared by the RLB Specialist Panel based on real practical AI usage in the workflows the respective audience uses AI for. For Regulation 1.44, the Panel set a deliverable-style operational brief: prepare guidance for a margin team configuring system parameters for separate account compliance, specifying which currencies require same-day collection, which receive extended deadlines, and the exact timing requirements for each currency category. Two frontier AI subjects answered with web search enabled, mirroring how associates and counsel at law firms actually use AI assistants on a finalised rule under client time pressure.

The Panel binds each AI finding to verbatim regulator-issued source text held as primary substrate, in this case the eCFR text of 17 CFR Section 1.44 as of June 2026.

What the models got wrong

Claude Opus 4.7, with web search active, wrote: "For margin called in a fiat currency other than USD, the deadline may be extended by up to one additional business day... T+1 should be the default presumption... For all other non-USD currencies, default to same-day (T)." The structural failure is the assignment of every Appendix A currency to a T+1 deadline. Section 1.44(f)(2) sets T+2 for that tier, "the end of the second business day after the day on which the margin call is issued." The model also assigned every remaining non-USD, non-Canadian-dollar currency to same-day collection.

Section 1.44(f)(3) grants those currencies T+1, end of the business day after the day on which the call was issued. The reconstruction matches the format of pre-finalisation third-party law-firm summaries and overrode the regulation's actual enumeration. Citation ID RLB-H-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001-Opus47 carries the full audit trail.

Claude Sonnet 4.6, given the same brief, wrote: "TIER 1 EXTENSION, OTHER NON-USD/CAD FIAT CURRENCIES (T+1, 12:00 p.m. ET) Deadline: 12:00 p.m. ET on the FIRST U.S. business day after the business day on which the margin call was issued." Section 1.44(f)(3) sets no intraday cutoff. The "12:00 p.m. ET" figure does not appear in the rule, in Appendix A, or in the CFTC release.

A separate Specialist Panel re-probe surfaced a related Sonnet 4.6 issue on the same rule: the model cited a URL on the Sidley Austin website to support a two-tier reading; the Panel verified the cited page does not exist. The Specialist Panel flagged the cited URL as Fabricated. Citation ID RLB-H-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001-Sonnet46 carries the full audit trail.

Why this matters for law firm legal teams

The exposure is not theoretical. The 2-page client memo is the firm's primary deliverable on a question of this kind, and an associate working under partner deadline pressure will treat an AI-drafted skeleton as the starting point. A memo built off the Opus 4.7 reconstruction would advise the FCM client to call Appendix A margin a full business day earlier than the rule permits, an error the client may institutionalise as policy before the firm's review caught it.

A memo built off the Sonnet 4.6 output would document a noon Eastern Time cutoff as a regulatory deadline, citing the firm's authority to a standard the CFTC has not set. Both errors travel: into the partner briefing pulled from the memo, into the board summary derived from the briefing, into the comparison table seeded into the firm's knowledge base. An examination response prepared off any of those derived deliverables will inherit the same error. And opposing counsel in a margin dispute will look at the documented firm position before they look at the rule.

The regulator's actual position

Section 1.44(f) is explicit on three tiers. Tier 1, U.S. and Canadian dollars, close of the Fedwire Funds Service on the same business day the margin call is issued. Tier 2, the ten currencies listed in Appendix A (AUD, CNY, HKD, HUF, ILS, JPY, NZD, SGD, ZAR, TRY), end of the second business day after the day on which the margin call is issued (T+2). Tier 3, "fiat currencies other than U.S.

Dollars, Canadian Dollars, or currencies listed in appendix A to this part shall be considered in compliance with the requirements of this paragraph (f) if received by the applicable futures commission merchant no later than the end of the business day after the day on which the margin call is issued" (T+1). No intraday cutoff appears anywhere in Section 1.44(f). The Tier 2 membership list is regulator-defined and not derivable from the currencies' intuitive groupings.

What this tells us about AI for law firm legal teams

The taxonomy lens for the firm is the same lens the Panel applies across recently finalised CFTC and prudential rules: Enumeration Collapse, the substitution of a plausible structure for the regulation's actual enumeration, presented as a usable deliverable with no surface signal to the reader. Both Regulation 1.44 outputs carried checklist formatting, jurisdictional groupings, internal cross-references, and no hedging language. The error is not visible at runtime; it is only catchable against the primary text.

For law firms, the work-product implication is structural: any AI-assisted deliverable that touches a recently finalised regulation's enumeration, currency lists, cessation triggers, threshold buckets, deadline tiers, requires verification against the primary text before the deliverable leaves the firm. A reasonableness check against the AI's own output is not sufficient.

What the RLB Specialist Panel is doing about it

The RLB Specialist Panel documents both findings under the immutable citation IDs above, binds each finding to the verbatim Section 1.44(f) text it contradicts, and records the failure mode classifications (outdated for the Opus 4.7 finding, inference_drift for the Sonnet 4.6 finding) against the eCFR primary substrate document. The Panel records the methodology categories used to surface the failure so firms and AI labs can identify Enumeration Collapse and Fabricated Intraday Cutoff in their own workflows.

Partnership inquiries from law firms that want to test their own internal AI-assisted regulatory drafting against the same probe family, and from AI labs that want to test their models against the same brief, are received at the Specialist Panel level.

What law firm legal teams should do

Firm-level actions, in order of work-product impact:

• Treat any AI-assisted Regulation 1.44 deliverable as draft until verified against 17 CFR Section 1.44(f) and Appendix A directly. The verbatim text, not a summary.
• Flag intraday Eastern Time cutoffs in any AI output on the T+1 tier as a diagnostic of reconstructed (not retrieved) text. No such cutoff exists in the rule.
• Name-check every currency in an AI-generated Tier 2 list against the Appendix A enumeration. Substitutions and omissions are the failure pattern.
• Verify any URL the AI cites as authority. Fabricated URLs to known law-firm domains have been observed on this rule and would be reproduced in firm deliverables unverified.
• Version the firm's internal knowledge-base entries on Regulation 1.44 and re-verify any entry generated since AI assistance was introduced into the drafting workflow.
• For partner-reviewed deliverables, route AI-assisted drafts through a primary-text verification step that is logged in the matter file, so a future correction does not require recall of unattributed work product.

Right of Reply

These findings and associated work have been put up in public with a view of the greater good for the development of a safer AI ecosystem. Any party reading this or any finding on reglegbrief.com may contact us and have an unconditional right of reply; the Specialist Panel will publish any factual correction or contextual response alongside the original finding, with no editorial gatekeeping. Researchers, regulators, and compliance teams with questions on methodology or specific findings can reach the Specialist Panel via the same channel.

Source & Methodology Standards

RegLeg Brief is operated by Verdus Technologies Pte. Ltd. (UEN 201616982R), incorporated in Singapore. The RLB Specialist Panel, with an aggregate of over 60 years of public-policy and industry experience, documents only confirmed hallucination findings, under a methodology that requires a verbatim regulator excerpt for every documented claim. All findings, citation IDs, model outputs, regulator excerpts, and methodology notes are open-access.

Primary source verified: 17 CFR § 1.44, Risk-based Capital Requirements and Margin Adequacy for Separate Accounts · Substrate documents: R2-REGULATION-17CFR_1_44_eCFR_asof_2026-06-04.pdf · eCFR: ecfr.gov · CFTC: cftc.gov

Citation IDs referenced:

• RLB-H-US-CFTC-FCM-MARGIN-ADEQUACY-SEPARATE-ACCOUNTS-REG-1-44-Q001-Opus47