ChatGPT, cited by the Chicago Housing Authority's counsel. Conceded as AI-produced.
After a six-week trial, a jury returned a $24 million verdict against the Chicago Housing Authority for two children who suffered lead poisoning and brain damage in a CHA apartment. CHA, represented by Goldberg Segalla, filed a post-trial motion arguing the court wrongly barred evidence of other lead sources. At page 25 it cited Mack v. Anderson, 441 Ill. App. 3d 819 (2021), claiming the "Supreme Court allowed evidence of other sources of environmental contamination" to challenge causation.
Plaintiffs responded that Mack v. Anderson "does not exist" and "appears to have been invented out of thin-air." CHA conceded the citation was improper. A partner, Danielle Malaty, admitted she used ChatGPT to research and draft the motion and did not verify the case before including it and its fictitious holding. A second nonexistent case, First Chicago Bank v. Brandwein, was also traced to her AI use.
Plaintiffs' sanctions motion then showed the problem was wider: fourteen cases in the post-trial motion were cited for propositions they did not support, with fabricated holdings and quotes. Signing partner Larry Mason insisted three times that only one case was at issue.
Offered for: admitting alternative-source evidence
The court granted the sanctions motion in part. Under Illinois Supreme Court Rule 137 it sanctioned Mason, the signer, $10,000 for filing three motions without reasonable inquiry and for repeatedly and falsely claiming only one citation was wrong. Goldberg Segalla, which accepted responsibility, was sanctioned $49,500 in plaintiffs' fees, and two sections of the post-trial motion were stricken. The court denied sanctions against Malaty and partners William O'Connell and Daniel Woods, none of whom signed the filings, noting Malaty had already lost her job and been sanctioned in a separate ChatGPT case.
Kingsfield returned REJECT. A nonexistent case offered as controlling authority. Counsel conceded it was produced using AI.
This matter is among the US case-law cases we ran Kingsfield against in the Charlotin benchmark. Kingsfield's existence check confirms every citation against the case-law corpus and an independent source (CourtListener) before any claim resting on it can stand. Across the run Kingsfield ruled 696 of 700 testable claims correctly, including every one of the 504 fabricated citations and 92 fabricated quotes.
Honest scope. This measures Kingsfield's ruling layer given an extracted citation, not an end-to-end run over a full brief. The case record and the AI-tool attribution come from Damien Charlotin's public database and the court order; the verdict is Kingsfield's. See the full 700-claim benchmark.