Sanction teardown · CA Texas, USA · 2026-09-24
Ex parte Bradley Thomas
What happened
In CA Texas, USA, a filing relied on an unnamed/unconfirmed AI tool to help draft legal argument. The court identified the following problems with the citations in that filing:
-
Fabricated (Case Law)Thomas cited Ex parte Arjona for a rule requiring an evidentiary hearing in scientifically complex habeas cases. The court found that no such case exists in the South Western Reporter; the cited volume and page instead correspond to In re E.C.R., a parental-rights case.
-
Fabricated (Case Law)Thomas cited Ex parte Rodriguez for propositions concerning habeas hearings, subpoenas, continuances, and legal standards. The court found the case apparently fictitious or inaccurately cited; the citation led to Winningham v. State, which concerned an unrelated sufficiency-of-the-evidence appeal.
-
Fabricated (Case Law)Thomas listed Garcia v. State as authority. The court found the case apparently fictitious or inaccurately cited; the cited volume and page instead corresponded to Vernon v. State.
-
False Quotes (Case Law)Thomas misquoted Hooper v. State. The court identified the misquotation as one of several briefing deficiencies and stated that the cited authority did not support the brief's presentation.
-
Misrepresented (Doctrinal Work)Thomas attributed to Strickland v. Washington the proposition that a trial strategy based on a misinterpretation of fundamental criminal law is objectively unreasonable per se. The court stated that this legal proposition does not appear anywhere in Strickland and that Thomas did not explain its application.
-
Misrepresented (Doctrinal Work)Thomas relied on Laster v. State for a proposition that appellate courts are constitutionally forbidden from protecting a verdict requiring an invented physical narrative. The court explained that Laster only sets out legal- and factual-sufficiency standards and does not support that proposition.
Which AI tool
an unnamed/unconfirmed AI tool. 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
Appeal affirmed; all complaints forfeited for inadequate briefing
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/3193/Bradley_Wayne_Thomas_v._State_USA_25-09-2026.pdf, via Damien Charlotin's public AI Hallucination Cases Database (CC0).