If a litigant or attorney submits to the court an AI-fabricated "fake precedent" without verifying it, a plan is being pursued to make them bear all litigation expense or to discipline the attorney.
The National Court Administration under the Supreme Court said on the 31st that the "AI-based false claims and evidence submission response task force (TF)" had prepared response measures for citations of false laws created with AI. The TF is composed of eight judges and two attorneys, among others, and has been in operation since Nov. last year.
Courts at all levels have reported cases in which litigants or attorneys cite and present false statutes or precedents generated by AI. The National Court Administration added a "verify false case number" function to the Judicial Information Disclosure Portal (portal.scourt.go.kr/) on the 20th of last month. By entering a case number, one can check whether the relevant judgment exists.
If a litigant or attorney cites false statutes created using AI and thereby incurs unnecessary litigation expense or delays the proceedings, the bench may order them to bear all or part of the resulting litigation expense.
In addition, for briefs that cite false statutes or precedents, the court may limit such statements at oral argument and may specify in the judgment that the related content is false. If an attorney submits AI-generated false statutes or precedents without proper verification, the bench may refer the attorney to the Korean Bar Association (KBA) for discipline.
The TF also proposed amending procedural laws to allow imposing fines when litigants and others cite false statutes. It proposed amending the Civil Procedure Rules to require parties using AI to notify opposing parties and the court of that fact and to verify whether key content is accurate.
An official at the National Court Administration said, "The expansion of Generative AI use is an irresistible trend of the times, and the courts need to prepare systematically for the changes and risks."