As artificial intelligence becomes increasingly embedded in the legal profession, courts around the world are confronting a growing problem: lawyers submitting briefs containing cases that do not exist.
What began as an isolated incident in the United States has evolved into a global concern, with judges, regulators, and disciplinary bodies warning attorneys that reliance on generative AI tools does not excuse failures to verify legal authorities.
The issue resurfaced recently when an Australian lawyer became one of the first in the country to face disciplinary action for filing AI-generated citations that could not be verified. The case has drawn renewed attention to similar incidents across multiple jurisdictions.
Here are six high-profile examples of lawyers finding themselves in trouble after courts discovered fabricated legal authorities in their filings.
1. The Avianca Case That Put AI Hallucinations on the Legal Map
The most famous AI-citation scandal occurred in 2023 in Mata v. Avianca Inc., a federal lawsuit in New York.
Attorney Steven Schwartz used ChatGPT to assist with legal research and subsequently submitted a brief containing six cases that were entirely fictitious. When opposing counsel and the court attempted to locate the cited authorities, they discovered the cases did not exist.
Even more remarkably, ChatGPT reportedly assured the attorney that the cases were genuine when asked to verify them.
U.S. District Judge P. Kevin Castel sanctioned the attorneys involved and imposed a $5,000 fine, emphasizing that lawyers have a duty to independently verify all legal authorities before presenting them to a court.
The case became a cautionary tale taught in law schools and legal ethics seminars worldwide.
2. Michael Cohen’s Brief Contained Nonexistent Cases
Former attorney and Trump associate Michael Cohen also found himself at the center of an AI-related controversy in 2023.
Court filings revealed that Cohen provided his attorney with legal citations generated by Google’s Bard AI platform. Several of the authorities referenced in the filing did not exist.
Cohen later explained that he mistakenly believed Bard functioned as a legal search engine rather than a generative AI system capable of fabricating information.
The incident highlighted how even non-lawyers participating in legal proceedings can inadvertently introduce AI-generated misinformation into court records.
3. U.S. Appeals Court Sanctions Attorneys Over AI Hallucinations
In 2026, a federal appeals court sanctioned two attorneys after finding that their filings contained fictitious legal authorities generated through AI.
The court rejected attempts to characterize the errors as simple mistakes and instead focused on the lawyers’ obligation of candor toward the tribunal.
Judges stressed that attorneys cannot delegate professional judgment to artificial intelligence systems and remain fully responsible for every citation submitted under their names.
The ruling reinforced a growing judicial consensus that AI-generated research must be rigorously checked before being filed.
4. UK Solicitor Criticized for Filing Fictional Authorities
The problem has not been confined to North America.
In the United Kingdom, a solicitor working for international law firm Pinsent Masons came under scrutiny after submitting an insolvency application that included legal authorities later found to be fictional.
The High Court criticized the filing, and the firm publicly apologized before self-reporting the matter to regulators.
The case became one of the highest-profile examples of AI-related citation failures within the UK legal system.
5. Regulators Investigate Growing Number of AI Citation Cases
Legal regulators in England and Wales have warned that fabricated AI-generated authorities are becoming an increasingly common issue.
Reports indicate that disciplinary bodies have examined multiple instances involving lawyers who relied on AI-generated research without adequate verification.
One reported case allegedly involved dozens of fake authorities appearing in appellate submissions.
The trend has prompted renewed discussions about legal ethics, technological competence, and the need for updated professional standards governing AI use.
6. Australian Lawyer Penalized in a Landmark First
The latest case emerged in Australia, where a Victorian solicitor became one of the first lawyers in the country to be formally disciplined over AI-generated legal citations.
The lawyer admitted filing authorities that could not be verified and acknowledged failing to properly check the research before submitting it.
The consequences were significant. Regulatory authorities restricted the solicitor’s ability to practice independently, prohibited the handling of trust money, and imposed a period of professional supervision.
The case has been widely described as a landmark moment for Australian legal ethics in the age of artificial intelligence.
A Growing Warning for the Legal Profession
These cases share a common theme: courts are making it clear that AI is not a defense for inaccurate legal filings.
Generative AI tools can draft documents, summarize research, and assist with analysis, but they can also produce convincing yet entirely fabricated information—a phenomenon often referred to as “hallucination.”
For judges and regulators, the responsibility remains unchanged.
Whether a citation comes from a law library, a junior associate, or an artificial intelligence platform, attorneys are ultimately responsible for ensuring that every authority cited in court is real, accurate, and relevant.
As AI adoption accelerates throughout the legal industry, these cases serve as a reminder that technological innovation does not replace professional responsibility—and that failing to verify AI-generated research can carry serious ethical and disciplinary consequences.
