A high-profile US law firm representing Alabama’s prison system is under scrutiny after one of its attorneys used ChatGPT to prepare legal documents that included fabricated case citations.
A federal judge in Birmingham, Michigan, is now considering sanctions against the firm, highlighting growing concerns over the unverified use of artificial intelligence in legal proceedings.
During a court hearing, Judge Anna Manasco revealed that one of the firm’s attorneys submitted two legal memoranda containing references to non-existent case law.
The lawyer, identified as Matt Reeves, admitted to using ChatGPT to assist with legal research but failed to verify the information before including it in official court documents.
Judge Manasco expressed concern over the issue, noting that courts across the country have already issued warnings about relying on AI-generated content for legal filings due to the risk of misinformation. She said she is reviewing a range of potential penalties, including financial sanctions, and has given the firm ten days to respond with a formal legal brief.
The documents in question, signed by four attorneys including Reeves, contained what are referred to as AI “hallucination.”
During the hearing, lawyers from the firm Butler Snow repeatedly apologized, acknowledging that the use of ChatGPT without proper verification violated both sound judgment and the firm’s internal policies. In a statement submitted to the court, the firm admitted embarrassment over the incident and reiterated that there is no excuse for referencing AI-generated legal precedent without independently confirming the sources.
Reeves took full responsibility for the erroneous citations, telling the judge that the decision was his alone and that his colleagues should not be penalized for the mistake.