What measures did Korean judges say are needed to address AI-generated fake legal citations?

Nearly 92 percent of respondents said institutional measures were needed. Suggestions included legal grounds for fines, verification requirements and disciplinary measures for lawyers who repeatedly cite false information.

How often have Korean judges themselves cited false legal precedents?

A total of 164 judges said they had cited false precedents themselves. The survey also found that 136 judges, or 57.6 percent, had encountered lawyers representing parties citing false precedents or other information.

How can AI-generated fake precedents affect court proceedings in South Korea?

A National Court Administration official said citations of false or nonexistent precedents may delay court proceedings. The official said effective countermeasures should be developed based on the survey results.

About 80% of Korean judges encounter AI-generated fake legal precedents: survey

Listen

Summary

The Supreme Court said Monday that Korean judges have encountered AI-generated false or nonexistent legal precedents cited in court cases. In a survey of 360 judges conducted from Oct. 6-16, 80.7 percent said they had seen nonexistent precedents cited, and 65.4 percent had encountered false precedents. Nearly 92 percent of respondents said institutional measures were needed. A National Court Administration official warned that such citations could delay proceedings.

Key Facts

The survey covered judges handling civil, family and administrative cases.
A total of 164 judges said they had themselves cited false precedents.
Some 136 judges, or 57.6 percent, said lawyers representing parties had cited false precedents or other information.
Forty-one judges, or 16.9 percent, had encountered citations of false or nonexistent laws and regulations.
Ten judges, or 4.1 percent, had seen citations of false or nonexistent theses and publications.

By Yonhap

Published Nov 2, 2026 12:27 pm KST

Eight in 10 Korean judges have experienced court cases in which false or nonexistent legal precedents generated by artificial intelligence (AI) were cited, a recent survey by the Supreme Court released Monday showed.

The top court published the outcome of the survey conducted on 360 judges from Oct. 6-16 as it sought to examine increasingly worrisome cases of "AI hallucination" — a term referring to a situation where AI models present false information as if they were true.

In the survey, about 80 percent of the judges in charge of civil, family and administrative cases said that they had experienced cases in which nonexistent or false precedents were cited.

Of those, 196 judges, or 80.7 percent, said they had seen the citations of nonexistent legal precedents in court cases, while 159 judges, or 65.4 percent, said have they encountered cases involving the citations of false precedents.

The survey also showed that 41 judges, or 16.9 percent, encountered cases in which false or non-existent laws and regulations were cited, while 10 judges, or 4.1 percent, said they have seen citations of false or non-existent theses and publications.

A total of 164 judges said they themselves have cited false precedents, while 136, or 57.6 percent, said that lawyers representing parties in court had cited false precedents or other information.

Nearly 92 percent of respondents said there need to be institutional measures to address the citations of AI-generated fake precedents.

Those measures include setting up legal grounds for imposing fines, establishing verification requirements and devising disciplinary measures against lawyers repeatedly citing false information.

An official at the National Court Administration pointed out that citing AI-generated false or nonexistent legal precedents may lead to a delay in court proceedings, underscoring a need to craft effective countermeasures based on the survey results.

Explore More