AI Labs · published 2026-05-26 · methodology v2.1

Claude Opus 4.7 with web search

RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q020-Opus47

What the RLB Specialist Panel found

Finding 20. Claude Opus 4.7 with web search

Question (paraphrased to protect IP): Which FCA Dear CEO letters remain in force after Consumer Duty implementation? Has the FCA withdrawn any letters that previously set consumer protection expectations?
AI's response: > "in April 2025 it retired around 90+ Dear CEO/portfolio letters, and in August 2025 retired 100+ multi-firm and thematic reports"
Regulator's text: From FS25/2 (March 2025): FCA reviewed all Dear CEO letters pre-dating the 2022-25 strategy and withdrew 90+ such letters. These are no longer live supervisory expectations. The multi-firm reports withdrawal was also announced in the same March 2025 document.
Why the AI went wrong: As with the earlier finding on this topic, the model again fabricated a two-event timeline — attributing the Dear CEO letter withdrawal to April 2025 and the multi-firm report withdrawal to August 2025 — when both were announced together in FS25/2 in March 2025. The recurrence of the same invented dates across two distinct questions suggests the model has constructed a persistent but incorrect internal account of this regulatory action, which it reproduces consistently regardless of how the question is framed.
Cited source(s):
https://www.fca.org.uk/news/news-stories/fca-simplifies-supervisory-letters — Pretextual
https://www.fca.org.uk/publication/correspondence/consumer-duty-letter-retail... — Pretextual
https://www.fca.org.uk/publication/correspondence/dear-ceo-letter-implementin... — Pretextual
Regulator portal (if any cited link is dud): https://www.fca.org.uk

Impact for this audience

The recurrence of the same fabricated April/August 2025 timeline (also present in Finding 13) across two differently-framed questions on the same topic confirms this is a persistent internal model representation rather than a random generation error. This has implications for how the model handles regulatory withdrawal and amendment records: it appears to have constructed a specific (incorrect) account of the FS25/2 action that it reproduces consistently. Correction pairs targeting this specific publication and its contents would be the most direct remediation.

References — raw findings (per AI model)

RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q020-Opus47 Claude Opus 4.7 (web search on)
RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q020-Sonnet46 Claude Sonnet 4.6 (web search on)

This finding also affects

← Previous finding Finding 7. Claude Opus 4.7 with web search Next finding → Finding 9. Claude Sonnet 4.6 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-Q020-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-v1-020--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-Q020-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-v1-020--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-Q020-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-v1-020--opus-47-websearch/.

BibTeX

@misc{reglegbrief_RLB_H_GB_FCA_CONSUMER_DUTY_PS22_9_Q020_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-Q020-Opus47},
  url       = {https://reglegbrief.com/regulators/j3/gb/fca/consumer-duty-ps22-9/whitepaper/finding/GB-FCA-GB-001-CONSUMER-DUTY-PS22-9-v1-020--opus-47-websearch/}
}

← Back to case study summary Case study detail →

About Citation IDs →