Submit incident
Documented

Four Phantom Cases Got a D.C. Appellate Brief Struck and an Attorney Referred for Discipline

August 12, 2025
Curated by Team Raidu · Reviewed by Shiva Ganesh
aiid:1682View source ↗
LinkedInX

What happened

In Douglas v. Deutsche Bank National Trust Co., the D.C. Court of Appeals struck Deutsche Bank's appellate brief in August 2025 after finding that four of its cited cases did not exist. The attorney who filed it, Loishirl W. Hall, acknowledged using Google's generative AI search tool to locate case authority and failing to verify those citations before the brief went in.

AI-assisted legal research tools surface case law by generating summaries and references from natural language queries. What they can also do, as this case made plain, is produce plausible-looking citations for cases that never happened. Hall's brief included four such citations, presented to an appellate court as real precedent, without a single check against an authoritative legal database to confirm they existed.

The court's response was unambiguous. It found the filing had misdirected the court and burdened judicial resources. Deutsche Bank's brief was struck entirely, a significant procedural sanction in any appellate litigation. The matter was also referred for disciplinary review, meaning Hall faces professional consequences beyond the immediate loss of the filing.

This is not the first case of its kind. Courts across the United States have sanctioned attorneys for filing AI-fabricated citations since at least 2023, when a federal judge in New York penalized lawyers who submitted a brief full of generated case references. The pattern is consistent: an attorney uses an AI tool to surface authority, trusts the output without verification, and files a document that cites cases the opposing party and the court cannot locate because they do not exist. The D.C. case is notable because the tool in question was Google's generative AI search feature rather than a standalone legal AI product, which reflects how broadly this capability has spread into ordinary research workflows.

What each of these cases shares is the absence of any mechanism that would have caught the failure before it reached a judge. A citation is either real or it is not, and confirming that takes minutes against a verified legal database. The gap here is not technical but procedural: no step in the workflow required proof that the sources existed before they were filed. A provable record of what a system returned, what a practitioner checked, and when each step occurred would close the window between a generated output and a court filing. Without that record, the verification burden falls entirely on individual professional judgment, and when it lapses, the consequences land on the court, the client, and the attorney's standing.

Reported impact

Affected parties
Not publicly disclosed
Harm type
Not publicly disclosed
Scale
Not publicly disclosed
Financial impact
Not publicly disclosed
Regulatory action
Not publicly disclosed

Classification

Organization
Not publicly disclosed
AI system
Not publicly disclosed
Industry
Not publicly disclosed
Country
Not publicly disclosed
Provider
Not publicly disclosed
Incident type
Not publicly disclosed

Relevant governance controls

Governance control mapping is not available for this record.

  • No controls mappedNot publicly disclosed

Control mapping is analytical. It does not state that any control would have prevented the incident.

Sources and evidence

This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.

AI Incident Database
Also catalogued in
Four Phantom Cases Got a D.C. Appellate Brief Struck and an Attorney Referred for Discipline
2025-08-12