Courts Sanction Lawyers Over AI-Generated Errors in Filings
A U.S. district court fined one attorney $1,000 and ordered training after fabricated quotes appeared in an AI-assisted brief. Similar sanctions have been issued in other U.S. cases as courts address the issue.
A U.S. district court sanctioned a lawyer last month after discovering seven fabricated or misrepresented quotes in a motion filed in a personal injury case. The attorney admitted using an artificial intelligence chatbot to draft the document and then submitting a revised version without review.
The court ordered the lawyer to pay a $1,000 fine and complete a three-hour course on AI-assisted legal practice. The judge stated that ignorance of AI risks is no longer an acceptable defense.
A database compiled since April 2023 now lists 1,600 examples of AI-generated errors in court filings from 35 countries. The United States accounts for 1,116 of those cases, followed by Canada with 173 and Australia with 74. The errors include invented case citations, plagiarized arguments, and references to nonexistent rulings.
Most involve mainstream chatbots used for research or text generation.
Judges have warned that such filings waste court resources and can damage the credibility of attorneys and the judiciary. In one recent federal case, four lawyers were fined a total of $8,000 and two were barred from appearing in a district court for two years.
Lawyers remain responsible for verifying all citations and arguments generated by AI tools, according to members of the American Bar Association task force on law and artificial intelligence. Specialized legal AI products show fewer errors but are not immune from them.
