AI Hallucination ResearchAudiencesSectorsInternational / MultilateralCorporate BankingComplianceDetail › Finding
Corporate Banking × Compliance — International / Multilateral · updated 2026-06-04
Share / Print Twitter LinkedIn Email

Finding#2 — Misidentified CPMI pre-validation implementation partner

RLB Citation ID: RLB-F-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007
AI's failure:Misstated Rule Risk for Corporate Banking × Compliance:Wrong deliverable
What the RLB Specialist Panel found
For Claude Opus 4.7 (web search on)
Question (paraphrased to protect IP)

An advisor to a South African financial institution needs to know which central banks are actively involved in piloting or implementing specific API harmonisation recommendations, and whether any central bank is specifically named as partnering with CPMI on the payment pre-validation API recommendation.

RLB's analysis

The model hedged toward the correct institution but could not confirm the explicit CPMI-SARB named partnership — a partnership announced in a November 2025 brief that appears to fall outside the retrieval pipeline's effective indexing. To fill the gap, the model cited a fabricated Bank of England URL as supporting evidence, producing a hallucinated citation in place of the accessible regulator record.

AI Head's analysis — what weakness in the AI model caused this

The fabricated Bank of England URL generated to fill the citation gap is the key signal: the model's citation-generation subsystem produced a plausible-looking but non-existent URL rather than returning 'no source found.' This points to a citation-generation pipeline that is not gated on verified retrieval — it will produce a URL-shaped output regardless of whether a real URL was retrieved.

Cited source(s)
  • https://www.bankofengland.co.za/news/2024/december/the-boe-welcomes-api-panel... — Fabricated
For Claude Sonnet 4.6 (web search on)
Question (paraphrased to protect IP)

An advisor was asked which central bank CPMI had specifically named as its partner for advancing the payment pre-validation API recommendation under d224. The AI responded that no central bank had been publicly identified in this role and proposed the Bank of England as the closest named involvement — when CPMI Brief No. 9 explicitly names the South African Reserve Bank (SARB) as CPMI's named collaboration partner conducting joint stakeholder interviews on the pre-validation API recommendation.

RLB's analysis

The model's retrieval pipeline did not surface CPMI Brief No. 9, which names SARB explicitly. Rather than returning a clean "not found," the model substituted the Bank of England — a higher-frequency institution in CPMI-adjacent content — as the most plausible named partner. The substitution follows a retrieval-ranking pattern where the model fills a retrieval gap with the highest-prior institution, displacing the correct lower-frequency answer.

AI Head's analysis — what weakness in the AI model caused this

The Bank of England substitution for SARB is a high-prior-institution fill pattern: the retrieval pipeline did not surface CPMI Brief No. 9, so the model substituted the most contextually plausible high-frequency institution. This failure mode will recur whenever the correct answer is a lower-frequency named institution in a recent sub-document that the retrieval index has not fully covered — a structural property of how institutional attribution fails under sparse indexing.

Impact for Compliance Teams in Corporate Banking Sector in international jurisdictions working with the Promoting the Harmonisation of Application Programming Interfaces to Enhance Cross-Border Payments: Recommendations and Toolkit

CPMI Brief No. 9 (November 2025) explicitly names the South African Reserve Bank (SARB) as CPMI's collaboration partner in advancing the payment pre-validation API recommendation. AI tools tested on this question either denied that any central bank is named as the specific partner or proposed the Bank of England as the closest analogue — and fabricated a supporting URL attributed to the Bank of England. A Compliance team briefing senior management or external counterparts on the CPMI's active implementation partnerships will propagate this misidentification into regulatory-engagement materials.

For Corporate Banking firms with correspondent relationships in South African or broader African markets, or those monitoring CPMI's live pilots for product or compliance positioning, the error is verifiable against publicly available CPMI publications — a reputational and credibility risk if it surfaces in a regulator-facing context.

References — raw findings (per AI model)
This finding also affects
← Previous finding Finding#1 — Fabricated self-assessment toolkit structure Next finding → Finding#3 — Invented per-recommendation stakeholder targeting
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.

RLB Citation ID: RLB-F-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007
Plain text Download
RegLeg Specialist Panel (2026). "Finding#2 — Misidentified CPMI pre-validation implementation partner — Corporate Banking × Compliance — International / Multilateral." Citation ID: RLB-F-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007. RegLegBrief AI Hallucination Research, published 2026-06-04. https://reglegbrief.com/regulators/j1/int/bis-cpmi/cpmi-api-harmonisation-cross-border-2024/sectors/corporate_banking/compliance/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007/
APA 7th edition Download
RegLeg Specialist Panel. (2026). Finding#2 — Misidentified CPMI pre-validation implementation partner [Hallucination finding RLB-F-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007]. RegLegBrief AI Hallucination Research. https://reglegbrief.com/regulators/j1/int/bis-cpmi/cpmi-api-harmonisation-cross-border-2024/sectors/corporate_banking/compliance/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007/
Bluebook / OSCOLA (US + UK legal) Download
RegLeg Specialist Panel, Finding#2 — Misidentified CPMI pre-validation implementation partner [RLB-F-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007], RegLegBrief AI Hallucination Research (June 04, 2026), https://reglegbrief.com/regulators/j1/int/bis-cpmi/cpmi-api-harmonisation-cross-border-2024/sectors/corporate_banking/compliance/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007/.
BibTeX Download
@misc{reglegbrief_RLB_F_INT_BIS_CPMI_API_HARMONISATION_CROSS_BORDER_2024_Q007,
  author    = {RegLeg Specialist Panel},
  title     = {Finding#2 — Misidentified CPMI pre-validation implementation partner},
  year      = {2026},
  publisher = {RegLegBrief AI Hallucination Research},
  note      = {Hallucination finding Citation ID: RLB-F-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007},
  url       = {https://reglegbrief.com/regulators/j1/int/bis-cpmi/cpmi-api-harmonisation-cross-border-2024/sectors/corporate_banking/compliance/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007/}
}
← Back to case study summary Case study detail →