1 AI-hallucination sanction decided between 2026-08-10 and 2026-08-16 (ISO week 2026-W33). Every ruling below is sourced from the public docket — case name, decided date, outcome, AI tool implicated (when named), and any monetary penalty language pulled verbatim from the ruling.
Week over week: 1 sanction filed the week of 2026-08-10 vs 14 sanctions the prior week (2026-08-03) — down 13 from the prior week.
One AI-hallucination ruling landed on the docket this week. Small-N weeks are the norm — the pattern is set by the trailing 12-week average, not by any individual week’s count.
Unidentified was the most-implicated AI tool this week.
The pattern each week is remarkably stable: existence failures (Layer 1) dominate. Attorneys file briefs containing citations that do not exist in any primary source — usually because a general-purpose LLM fabricated the citation and no deterministic pre-filing check caught it. Every ruling below is a preventable filing. See /verify for the pre-filing tool.
Federal Court of Australia · Decided 2026-08-11 · Court noted and rejected AI-generated erroneous quotations and citations, ignored the applicant's unsolicited post-hearing email, allowed summary judgment for the Respondent; no professional sanction or monetary penalty imposed. · AI tool: Unidentified.
Get the Sanctions Alert — a new issue every time a ruling lands.
Free double-opt-in newsletter. Case, sanction, fake cites, lesson. No promotional content.
Informational only. Not legal advice. Data derived from public court records; corrections at /contact.