AI Law - International Review of Artificial Intelligence LawCC BY-NC-SA Commercial Licence ISSN 3035-5451
G. Giappichelli Editore

03/11/2026 - Judges Seek Safeguards Against AI-Fabricated Court Citations (South Korea)

argument: Notizie/News - AI in Judicial Activities

Source: New Straits Times

South Korean judges have reported encountering AI-generated false or nonexistent legal precedents in court cases. The Supreme Court published findings from a survey conducted among 360 judges to examine concerns about AI hallucinations. The report describes hallucinations as instances in which AI presents incorrect information as factual. About eight in ten judges handling civil, family and administrative cases reportedly encountered citations to nonexistent or false precedents. The reported findings include 196 judges who encountered nonexistent precedents and 159 who encountered false precedents. Other findings include 41 judges encountering false or nonexistent laws and regulations, and ten encountering false or nonexistent academic works and publications.

The report also states that 164 judges acknowledged having cited false precedents themselves. Another 136 judges reported that lawyers representing parties had cited false precedents or other information. Nearly 92 per cent of respondents supported institutional measures addressing AI-generated fabricated precedents. Suggested responses included creating a legal basis for fines and introducing verification requirements. The proposed responses also included disciplinary measures against lawyers who repeatedly cite false information. A National Court Administration official warned that fabricated AI precedents could delay proceedings and called for effective countermeasures informed by the survey.