AI Hallucination ResearchAudiencesSectorsSingaporeInvestment BankingComplianceDetail › Finding
Investment Banking × Compliance — Singapore · published 2026-05-28 · methodology v2.1

Unverified description of Division 4, Part VI coverage

What the RLB Specialist Panel found

3. Unverified description of Division 4, Part VI coverage

  • Question (paraphrased to protect IP): What does Division 4 of Part VI of MAS Notice 637 cover?
  • Source regulation: MAS Notice 637 (Risk Based Capital Adequacy Requirements for Banks Incorporated in Singapore) (Regulator portal: https://www.mas.gov.sg)
  • What AI assistants typically say: The AI describes Division 4 of Part VI as covering submission requirements relating to capital instruments, while acknowledging within the same response that this characterisation could not be exhaustively verified from publicly available sources.
  • What the regulator actually says: MAS Notice 637 contains a detailed divisional structure within Part VI. The precise coverage of Division 4 is defined in the notice text as published and amended by MAS, and that text is the authoritative statement of the division's scope.
  • Why the AI went wrong: The AI appears to have inferred the division's content from the general subject matter of Part VI rather than retrieving verified text. The AI's own hedge — acknowledging that its answer could not be confirmed — reflects a recognised knowledge gap, yet the response still offered a substantive characterisation that a reader could act on. Presenting an unverified inference alongside a caveat does not make the inference safe to rely upon.
  • Cited source(s):
Impact for this audience

A Compliance team that relies on the AI's unverified description of Division 4, Part VI could draft internal capital policy or training materials that misstate the obligations in that division. If those materials influence how the firm manages the relevant capital instruments or submissions, the firm may be non-compliant with the actual notice requirements without being aware of it. Discovery during a MAS thematic review or supervisory inspection could result in findings against the firm's capital management framework, mandatory remediation, and — depending on the materiality of the misstatement — formal supervisory action.

References — raw findings (per AI model)
This finding also affects
← Previous finding Finding 2. Mischaracterisation of yellow-highlight convention in MAS amendment notices
Cite this finding

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.

Plain text
RegLeg Specialist Panel (2026). "Unverified description of Division 4, Part VI coverage — Investment Banking × Compliance — Singapore." Citation ID: . RegLegBrief AI Hallucination Research, published 2026-05-28. https://reglegbrief.com/audiences/sectors/sg/investment_banking/compliance/finding/q-NOTICE-637-CAPITAL-ADEQUACY-BANKS-2025-v1-022/
APA 7th edition
RegLeg Specialist Panel. (2026). Unverified description of Division 4, Part VI coverage [Hallucination finding ]. RegLegBrief AI Hallucination Research. https://reglegbrief.com/audiences/sectors/sg/investment_banking/compliance/finding/q-NOTICE-637-CAPITAL-ADEQUACY-BANKS-2025-v1-022/
Bluebook / OSCOLA (US + UK legal)
RegLeg Specialist Panel, Unverified description of Division 4, Part VI coverage [], RegLegBrief AI Hallucination Research (May 28, 2026), https://reglegbrief.com/audiences/sectors/sg/investment_banking/compliance/finding/q-NOTICE-637-CAPITAL-ADEQUACY-BANKS-2025-v1-022/.
BibTeX
@misc{reglegbrief_,
  author    = {RegLeg Specialist Panel},
  title     = {Unverified description of Division 4, Part VI coverage},
  year      = {2026},
  publisher = {RegLegBrief AI Hallucination Research},
  note      = {Hallucination finding Citation ID: },
  url       = {https://reglegbrief.com/audiences/sectors/sg/investment_banking/compliance/finding/q-NOTICE-637-CAPITAL-ADEQUACY-BANKS-2025-v1-022/}
}
← Back to case study summary Case study detail →