AI Hallucination Research › Regulators › Major advanced economies › GB › FCA › CONSUMER-DUTY-PS22-9 › White paper › Detail › 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-Q005-Opus47
What the RLB Specialist Panel found
  • Question (paraphrased to protect IP): Does the Consumer Duty apply to micro-enterprises or small charities? Are they 'retail customers' under PRIN 2A?
  • AI's response: > "charities with annual income under £1 million"
  • Regulator's text: Retail customer includes consumers (individual persons), micro-enterprises, and charities with annual turnover less than £1 million. Definition varies by sourcebook.
  • Why the AI went wrong: The FCA's definition uses "annual turnover" as the threshold measure for charities. The model substituted "annual income" — a different financial metric that would produce different results in practice for charities that have income sources distinct from trading turnover. This is a precise definitional substitution: the word swap is small but changes which charities fall within scope, and reflects the model applying a common synonym rather than retrieving the regulator's exact defined term.
  • Cited source(s):
  • https://handbook.fca.org.uk/handbook/PRIN/2A/1.html — Pretextual
  • https://www.fca.org.uk/publication/finalised-guidance/fg22-5.pdf — Pretextual
  • Regulator portal (if any cited link is dud): https://www.fca.org.uk
Impact for this audience

This finding is a precise defined-term substitution: 'annual income' for 'annual turnover'. This implicates the model's lexical handling of regulatory definitions — specifically, whether its training corpus contained the exact FCA definition or a paraphrase. The fact that web search did not surface the correct term suggests either the retrieval step did not reach the Handbook definition, or the model did not use the retrieved content to override its training inference. A targeted eval checking exact regulatory defined terms (particularly financial thresholds with close synonyms) would catch this class of error.

References — raw findings (per AI model)
This finding also affects
← Previous finding Finding 1. Claude Opus 4.7 with web search Next finding → Finding 3. 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-Q005-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-v1-005--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-Q005-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-v1-005--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-Q005-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-v1-005--opus-47-websearch/.
BibTeX 
@misc{reglegbrief_RLB_H_GB_FCA_CONSUMER_DUTY_PS22_9_Q005_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-Q005-Opus47},
  url       = {https://reglegbrief.com/regulators/j3/gb/fca/consumer-duty-ps22-9/whitepaper/finding/GB-FCA-GB-001-CONSUMER-DUTY-PS22-9-v1-005--opus-47-websearch/}
}
← Back to case study summary Case study detail →
About Citation IDs →