🔍 Read the full analysis: 360 Judges Consider The Impact Of AI On Fake Precedents In Korea on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
South Korea’s National Court Administration reported that about 80% of surveyed civil, family and administrative judges had encountered false or distorted citations. Judges called for institutional action, including fines, citation-verification duties and disclosure of AI use; the survey does not measure how often fake citations appear in filings overall.
South Korea’s National Court Administration said about 80% of surveyed judges handling civil, family and administrative cases had encountered false or distorted legal citations, including nonexistent court precedents, in filings. The findings, published November 2, 2026, came with calls from judges for formal measures such as fines, citation checks and disclosure of AI use.
The survey asked 360 judges from courts at all levels to report their experiences. Among judges who had encountered false citations, 196, or 80.7%, identified nonexistent precedents, while 159, or 65.4%, said real precedents had been cited with false content. Respondents could select more than one type of material.
The “about 80%” figure applies to responding judges in civil, family and administrative cases, not all 360 participants. The Korean daily Kyunghyang Shinmun reported that 243 of the 360 respondents, or 67.5% overall, had encountered fake citations. The court administration’s release does not give that overall figure, and its findings describe judges’ reported experiences rather than the share of court filings containing false material.
Judges said false material came from both litigants and legal representatives. 164 judges (69.5%) identified parties representing themselves, while 136 (57.6%) identified lawyers acting for a party. These were multiple-response answers. The administration’s release said false citations appeared in documents submitted by lawyers as well as by parties themselves.
The court administration described examples including fabricated judgments formatted to resemble Supreme Court decisions, case numbers attached to legal reasoning that could not be verified, and real rulings presented with altered or misleading meanings. It also cited filings that retained text from AI conversations or notes about litigation strategy.
Pressure for Rules on AI Citations
The findings put a practical burden on courts and opposing parties: each questionable citation may need to be checked before a judge can assess the legal argument. The National Court Administration said false and nonexistent precedents create unnecessary verification work for opposing parties and can impede the speedy progress of trials.
The survey also shows that judges’ concern is not limited to unrepresented litigants. Their responses included citations attributed to licensed lawyers, raising questions about professional review when legal documents use AI tools. The survey does not establish that AI caused every false citation, but its release specifically connects the issue to AI-generated errors and includes examples of chatbot text appearing in filings.
Most respondents favored an institutional response: 91.7% said one was needed, and 64.8% described it as very necessary. The most popular proposed measures were a legal basis for fines, a duty to verify cited precedents and a duty to disclose AI use. Those preferences indicate support among surveyed judges, not rules already in force.
Products Worth Considering
Logitech MX Mechanical Wireless Illuminated Keyboard, Clicky
Samsung 34-Inch Odyssey G5 Ultra-Wide Curved Gaming Monitor
Apple MacBook Neo 13-inch Laptop with A18 Pro Chip, 8GB Memory, 512GB SSD, and Touch ID – Blush
As an affiliate, we earn on qualifying purchases.
What the Court Survey Measured
The National Court Administration, the administrative arm of South Korea’s Supreme Court, conducted the survey from October 6 to 16, 2026, and published the results on November 2. The four-page release was issued by its Judicial Support Office and presented as a survey on citations of AI-generated false precedents and similar material.
The administration said AI “hallucinations” can lead to citations of false or nonexistent precedents. That is the court body’s explanation of the problem, not a finding that every reported instance came from AI. The survey asked judges what they had seen; it did not count or independently audit court filings.
The release also listed false or nonexistent statutes and academic materials, though less frequently than false precedents. 41 judges (16.9%) selected statutes and 10 (4.1%) selected papers or literature. Percentages in the release are based on respondents to each question, and respondents could give more than one answer.
There is also a translation issue in one English-language account. Yonhap’s English version, republished by The Korea Herald, described 164 judges as saying they themselves had cited false precedents. The Korean release instead says those judges identified the parties themselves as the source. The distinction matters: the figure does not report judges admitting to citing false material.
Limits of the Survey Evidence
The survey does not show how many filings contained false citations, how often the problem occurs across the court system, or what share of cases involved AI. The administration did not disclose how many judges were invited, and the release does not provide the exact wording of the questions. Its results capture what participating judges recalled encountering, rather than a statistical audit of court documents.
The figures also rely on different denominators. The 80.7% figure refers to 196 judges among those answering the relevant question about citation types, while the broader 80% description refers to judges handling civil, family and administrative cases. The separately reported 67.5% figure is an arithmetic based on all 360 respondents, as reported by Kyunghyang Shinmun. The survey cannot establish from these figures alone whether false citations are increasing or how many were generated by AI rather than other means.
It is also not yet clear what safeguards would apply if new rules are adopted, how any duty to verify or disclose AI use would be enforced, or whether fines would be available in both civil and criminal proceedings. The survey records judges’ preferred responses; it does not announce that these measures have been enacted.
Proposed Procedural Law Changes
The National Court Administration plans to propose changes to South Korea’s Civil and Criminal Procedure Acts that would provide a legal basis for fines, according to the source report. The survey release’s figures show judges also favored requirements to verify cited precedents and disclose AI use, but the available material does not establish whether those duties will be included in a legislative proposal.
The next steps, including the timing and wording of any proposed amendments, remain unclear. Until a proposal is made and any changes are adopted, the survey’s recommendations should be understood as reported preferences, not new obligations for lawyers or litigants. Further detail from the court administration or lawmakers would clarify the proposed scope, enforcement and safeguards.
Key Questions
What did South Korea’s judges report?
The National Court Administration said about 80% of responding judges in civil, family and administrative cases had encountered false or distorted citations. The survey included 360 judges across courts at all levels.
Does the survey show that 80% of court filings contain fake precedents?
No. The figure describes judges who said they had encountered such citations. The survey did not count filings or estimate the share of court documents containing false material.
Did the survey find that lawyers used false citations?
Judges identified both parties representing themselves and lawyers as sources. 136 respondents, or 57.6%, selected lawyers acting for a party; respondents could select more than one source.
What measures did judges support?
Most respondents called for an institutional response. Popular options included a legal basis for fines, a duty to verify citations and a duty to disclose AI use. These were survey preferences, not enacted requirements.
What happens next?
The National Court Administration plans to propose changes to the Civil and Criminal Procedure Acts to allow fines. The timing, details and final contents of any proposal have not been provided.
Source: ThorstenMeyerAI.com
