AI Hallucination ResearchRegulatorsMajor advanced economiesGBFCACONSUMER-DUTY-PS22-9White paperDetail › Finding
AI Labs · published 2026-05-26 · methodology v2.1

Claude Opus 4.7 with web search

RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q016-Opus47
What the RLB Specialist Panel found
  • Question (paraphrased to protect IP): What has the FCA said publicly about firms' Consumer Duty compliance in the first year of implementation? What specific concerns have senior FCA officials raised?
  • AI's response: > "Sheldon Mills (FCA Executive Director of Consumers and Competition) delivered speeches on 20 February 2024 ('Consumer Duty: the art of the possible in a year') and 31 July 2024 (one-year-on event). His themes/concerns have included: (i) closed products... (ii) fair value assessments... (iii) consumer understanding... (iv) data and outcomes monitoring."
  • Regulator's text: The FCA speech by Sheldon Mills at the Consumer Duty one-year event (July 2024) was not independently accessible for verbatim verification.
  • Why the AI went wrong: The model produced specific speech dates, a titled speech name, and a thematic breakdown of regulatory concerns without access to a verified source for the July 2024 speech content. While some elements may reflect the general direction of FCA communications in that period, the model presented fabricated specifics — including a precise speech title and date — with the same surface confidence as verifiable facts. When source content is not retrievable, this model tends to fill the gap with plausible-sounding constructed detail rather than disclosing the limitation.
  • Cited source(s):
  • https://www.fca.org.uk/news/speeches/consumer-duty-art-possible-year — Pretextual
  • Regulator portal (if any cited link is dud): https://www.fca.org.uk
Impact for this audience

This finding exposes a confidence-calibration failure on unverifiable content: the model produced specific speech dates and titles for FCA communications it could not have retrieved, presenting them with the same surface confidence as verified facts. This implicates the model's uncertainty signalling — specifically, when retrieved content is inaccessible, the model defaults to generating plausible-sounding specifics rather than disclosing the source gap. The citation step then appended a plausible-looking URL rather than flagging the absence of a verified source.

References — raw findings (per AI model)
This finding also affects
← Previous finding Finding 4. Claude Opus 4.7 with web search Next finding → Finding 6. Claude Opus 4.7 with web search
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). "Claude Opus 4.7 with web search — AI Labs." Citation ID: RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q016-Opus47. RegLegBrief AI Hallucination Research, published 2026-05-26. https://reglegbrief.com/regulators/j3/gb/fca/consumer-duty-ps22-9/whitepaper/finding/GB-FCA-GB-001-CONSUMER-DUTY-PS22-9-v1-016--opus-47-websearch/
APA 7th edition
RegLeg Specialist Panel. (2026). Claude Opus 4.7 with web search [Hallucination finding RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q016-Opus47]. RegLegBrief AI Hallucination Research. https://reglegbrief.com/regulators/j3/gb/fca/consumer-duty-ps22-9/whitepaper/finding/GB-FCA-GB-001-CONSUMER-DUTY-PS22-9-v1-016--opus-47-websearch/
Bluebook / OSCOLA (US + UK legal)
RegLeg Specialist Panel, Claude Opus 4.7 with web search [RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q016-Opus47], RegLegBrief AI Hallucination Research (May 26, 2026), https://reglegbrief.com/regulators/j3/gb/fca/consumer-duty-ps22-9/whitepaper/finding/GB-FCA-GB-001-CONSUMER-DUTY-PS22-9-v1-016--opus-47-websearch/.
BibTeX
@misc{reglegbrief_RLB_H_GB_FCA_CONSUMER_DUTY_PS22_9_Q016_Opus47,
  author    = {RegLeg Specialist Panel},
  title     = {Claude Opus 4.7 with web search},
  year      = {2026},
  publisher = {RegLegBrief AI Hallucination Research},
  note      = {Hallucination finding Citation ID: RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q016-Opus47},
  url       = {https://reglegbrief.com/regulators/j3/gb/fca/consumer-duty-ps22-9/whitepaper/finding/GB-FCA-GB-001-CONSUMER-DUTY-PS22-9-v1-016--opus-47-websearch/}
}
← Back to case study summary Case study detail →