Sanction teardown · Upper Tribunal, UK · 2025-09-29
ANPV & SAPV v Secretary of State for the Home Department
What happened
In Upper Tribunal, UK, a filing relied on Microsoft Copilot to help draft legal argument. The court identified the following problems with the citations in that filing:
-
Fabricated (Case Law)Grounds cited a Honduras country guidance decision that does not exist; judge found the MN Honduras CG [2016] authority fabricated.
-
Fabricated (Case Law)Grounds cited an Eritrea appellate decision that does not exist (YH (Eritrea) [2008] UKAIT 00070); judge accepted counsel's concession that case was fictional.
-
Misrepresented (Case Law)HK v SSHD [2006] EWCA Civ 1037 was cited as supporting 'trauma-informed principles' but judge found the case does not discuss that; citation misapplied.
-
Misrepresented (Case Law)MA (Somalia) v SSHD was referenced but counsel produced the Supreme Court decision instead of the Court of Appeal authority relied upon; judge found no support for the asserted proposition.
-
Misrepresented (Case Law)A citation labelled 'SC (Protection) [2020] UKSC 9' did not match the case title; judge found a citation mismatch and irrelevant authority was cited for internal relocation principles.
-
Misrepresented (Case Law)R (YH) v SSHD [2010] EWCA Civ 116 was deployed in grounds but judge noted ChatGPT has recently wrongly deployed that case in similar submissions; authority was misused.
Which AI tool
Microsoft Copilot. Note: Charlotin's public database records tool attribution only where a court order, brief, or reporting on the matter states it explicitly; "unidentified" or "implied" means the record indicates AI use but does not name a specific product — we do not guess.
Outcome
Show Cause Order
How Citation Safe would have caught this
Citation Safe runs three deterministic layers before a brief is filed: (1) does the citation exist against CourtListener's database of published opinions, (2) if quoted, does that exact language appear in the source, (3) does the cited case actually support the proposition it is cited for. Fabricated case citations fail Layer 1. Fabricated or misattributed quotations fail Layer 2 even when the underlying case is real. Misrepresented holdings — a real case cited for a proposition it does not support — are the target of Layer 3. None of these checks involve asking another language model whether the citation looks right; they are lookups and text-matches against the actual source, which is why a hallucinated citation has to survive a direct lookup against the authoritative source — not another model's opinion — to earn a VERIFIED stamp; our measured false-verify rate is published live at /quality.
Check a brief before you file it → · See our live false-verify rate
Source: https://www.damiencharlotin.com/documents/884/UI-2025-003373___UI-2025-003374.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).