AI Hallucination Research › Regulators › Global standard-setters › INT › BIS-CPMI › CPMI-API-HARMONISATION-CROSS-BORDER-2024 › White paper › Detail › Finding
AI Labs · Last updated 7 Jun 2026 · methodology vv2.3 · Hallucination Register

Finding on 'Q007 Probe' for Claude Opus 4.7 with web search ON

RLB Citation ID: RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007-Opus47
What the RLB Specialist Panel found

Finding on 'Q007 Probe' for Claude Opus 4.7 with 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.
  • AI's response: "it is plausible that SARB is engaged with the d224 implementation track, but I cannot point to a public CPMI statement naming SARB as the specific partner on the pre-validation API recommendation."
  • Regulator's text: CPMI Brief No. 9 (Nov 2025): 'The CPMI, in collaboration with the South African Reserve Bank (SARB), has been advancing the API recommendation on payment pre-validation by conducting interviews with market stakeholders.'
  • Why the AI went wrong: The model retrieved enough surface signal to identify SARB and the payment pre-validation recommendation but treated retrieval-uncertainty as truth-uncertainty, producing a hedge where CPMI Brief No. 9 had already removed the uncertainty by naming the partnership explicitly. The model's calibration here is biased toward hedge when it cannot directly verify a retrieved fact, even when the regulator has published the verifying source.
  • Cited source(s):
Impact for this audience

The model retrieved the SARB-CPMI pre-validation partnership context — it had enough signal to know the question was about a named central-bank collaboration on a specific recommendation — but downgraded a regulator-confirmed fact to speculative hedge phrasing ('plausible but unverified'). This is calibration drift in a direction that looks like safety: the model hedged where it should have committed. The training and retrieval pipeline appears to penalise commitment without verifying retrieval, producing under-confidence on facts the regulator has itself published.

For evaluators, the high-value probe is the asymmetry: when the AI has retrieved enough to identify the entity at issue (SARB, payment pre-validation, CPMI Brief No. 9), commitment behaviour should match retrieval depth rather than defaulting to hedge. A confidence-calibration eval that scores against regulator-published source text would surface this drift before it ships to production users.

References — raw findings (per AI model)
This finding also affects
Next finding → Finding on 'Q007 Probe' for Claude Sonnet 4.6 with web search ON
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-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007-Opus47
Plain text Download 
RegLeg Specialist Panel (2026). "Finding on 'Q007 Probe' for Claude Opus 4.7 with web search ON — AI Labs." Citation ID: RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007-Opus47. RegLegBrief AI Hallucination Research, published 2026-06-07. https://reglegbrief.com/regulators/j1/int/BIS-CPMI/CPMI-API-HARMONISATION-CROSS-BORDER-2024/whitepaper/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007--opus-47-websearch/
APA 7th edition Download 
RegLeg Specialist Panel. (2026). Finding on 'Q007 Probe' for Claude Opus 4.7 with web search ON [Hallucination finding RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007-Opus47]. RegLegBrief AI Hallucination Research. https://reglegbrief.com/regulators/j1/int/BIS-CPMI/CPMI-API-HARMONISATION-CROSS-BORDER-2024/whitepaper/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007--opus-47-websearch/
Bluebook / OSCOLA (US + UK legal) Download 
RegLeg Specialist Panel, Finding on 'Q007 Probe' for Claude Opus 4.7 with web search ON [RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007-Opus47], RegLegBrief AI Hallucination Research (June 07, 2026), https://reglegbrief.com/regulators/j1/int/BIS-CPMI/CPMI-API-HARMONISATION-CROSS-BORDER-2024/whitepaper/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007--opus-47-websearch/.
BibTeX Download 
@misc{reglegbrief_RLB_H_INT_BIS_CPMI_API_HARMONISATION_CROSS_BORDER_2024_Q007_Opus47,
  author    = {RegLeg Specialist Panel},
  title     = {Finding on 'Q007 Probe' for Claude Opus 4.7 with web search ON},
  year      = {2026},
  publisher = {RegLegBrief AI Hallucination Research},
  note      = {Hallucination finding Citation ID: RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q007-Opus47},
  url       = {https://reglegbrief.com/regulators/j1/int/BIS-CPMI/CPMI-API-HARMONISATION-CROSS-BORDER-2024/whitepaper/finding/INT-BIS-CPMI-INT-001-CPMI-API-HARMONISATION-CROSS-BORDER-2024-v1-007--opus-47-websearch/}
}
← Back to case study summary Case study detail →
About Citation IDs →

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.