An FCM that began accepting bitcoin, ether, and USDC as customer margin collateral in January 2026 under the CFTC digital asset pilot program wants to know what changes at the end of the initial three-month onboarding phase — specifically which obligations cease and which continue, and whether the weekly digital asset holdings reporting requirement and incident-reporting condition sunset or persist beyond the initial phase.
The model correctly identified that some initial-phase conditions sunset but incorrectly extended the sunset to the weekly reporting requirement, which the regulator explicitly preserved as a continuing obligation. The failure pattern is consistent with structural inference from the phased framework — the model appears to have generalised "initial-phase conditions" to cover all cadence-based requirements, rather than reading the regulator's specific enumeration of what does and does not lapse.
This finding implicates the model's handling of phased obligation structures where partial sunset language is present. The model correctly identified some sunset conditions but over-generalised the sunset to a continuing obligation — a pattern that suggests the training-data representation of this instrument came primarily from third-party summaries that flatten the obligation lifecycle rather than from the regulator's specific enumeration. Web search was active and did not correct the error, indicating the retrieval stack did not surface the primary text's carve-out language.
An FCM began accepting bitcoin, ether, and USDC as customer margin collateral in January 2026 under the CFTC's digital asset pilot. At the end of the initial three-month phase, does the weekly obligation to report total digital asset holdings in each customer account class cease or continue, and which other initial-phase conditions do sunset at that point?
The model fabricated a source — "March 2026 CFTC Staff FAQs" — to support an answer that is directly contradicted by the regulator's text, and presented the termination as a precisely-worded procedural rule. This is the more severe failure form: not merely an inference from the instrument's structure, but a confabulated authority cited to give the wrong answer the appearance of documentary grounding.
This is the most severe finding in this paper: the model fabricated a specific source document — 'March 2026 CFTC Staff FAQs' — to support an answer directly contradicted by the regulator's text, and presented the termination as a precisely-worded procedural rule. This implicates the calibration signal for named-source citations: the model committed to a document title and date without apparent retrieval basis rather than flagging uncertainty. It also implicates the training-data representation of the amendment cycle — the model's confident wrong answer suggests it is reconstructing from a plausible structural template, not retrieving the governing text.
A Risk team building the firm's FCM oversight framework from AI output on this question would set the reporting calendar to terminate weekly digital asset holdings submissions after the initial three-month phase — when the obligation in fact continues indefinitely. The internal policy, MI reporting pack, and compliance monitoring controls would all be calibrated to the wrong timeline.
When the error surfaces — under a CFTC examination or internal audit challenge asking the team to evidence its reporting cadence against the relief letter — the firm faces both a remediation cost (reconstructing the compliance record for the gap period) and the enforcement exposure that attaches to any ongoing failure to submit required reports to the CFTC. The CFTC's authority to impose civil monetary penalties and refer cases for criminal prosecution means this is not a low-stakes administrative correction.
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). "Finding#1 — Weekly reporting obligation sunset vs. continuation — Investment Banking × Risk — United States." Citation ID: RLB-F-US-CFTC-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-Q006. RegLegBrief AI Hallucination Research, published 2026-06-04. https://reglegbrief.com/regulators/j3/us/cftc/digital-asset-collateral-tokenized-assets-staff-guidance-2025/sectors/investment_banking/risk/finding/US-CFTC-US-001-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-v1-006/
RegLeg Specialist Panel. (2026). Finding#1 — Weekly reporting obligation sunset vs. continuation [Hallucination finding RLB-F-US-CFTC-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-Q006]. RegLegBrief AI Hallucination Research. https://reglegbrief.com/regulators/j3/us/cftc/digital-asset-collateral-tokenized-assets-staff-guidance-2025/sectors/investment_banking/risk/finding/US-CFTC-US-001-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-v1-006/
RegLeg Specialist Panel, Finding#1 — Weekly reporting obligation sunset vs. continuation [RLB-F-US-CFTC-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-Q006], RegLegBrief AI Hallucination Research (June 04, 2026), https://reglegbrief.com/regulators/j3/us/cftc/digital-asset-collateral-tokenized-assets-staff-guidance-2025/sectors/investment_banking/risk/finding/US-CFTC-US-001-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-v1-006/.
@misc{reglegbrief_RLB_F_US_CFTC_DIGITAL_ASSET_COLLATERAL_TOKENIZED_ASSETS_STAFF_GUIDANCE_2025_Q006,
author = {RegLeg Specialist Panel},
title = {Finding#1 — Weekly reporting obligation sunset vs. continuation},
year = {2026},
publisher = {RegLegBrief AI Hallucination Research},
note = {Hallucination finding Citation ID: RLB-F-US-CFTC-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-Q006},
url = {https://reglegbrief.com/regulators/j3/us/cftc/digital-asset-collateral-tokenized-assets-staff-guidance-2025/sectors/investment_banking/risk/finding/US-CFTC-US-001-DIGITAL-ASSET-COLLATERAL-TOKENIZED-ASSETS-STAFF-GUIDANCE-2025-v1-006/}
}