South Korean Judges Report Surge in AI-Generated Fabricated Court Citations

A recent survey conducted by the National Court Administration in South Korea has revealed a significant challenge facing the nation's legal system. More than eight out of ten judges handling civil, family, and administrative cases reported encountering citations of nonexistent or fabricated court rulings in legal documents. This trend has sparked urgent discussions regarding the impact of artificial intelligence on the integrity of judicial proceedings. The survey, which gathered responses from 360 judges across the country between October 6 and October 16, 2026, highlights the growing prevalence of AI-generated materials in litigation. According to the findings, 80.7 percent of respondents have seen citations of rulings that do not exist. Additionally, 65.4 percent of judges reported cases where rulings were falsely cited or manipulated to misrepresent their original intent. In some instances, litigants have submitted documents that mimic the formal structure of actual Supreme Court judgments. Other filings included legal reasoning paired with fabricated case numbers that judges could not verify through official databases. Some submissions even contained accidental traces of AI interaction, such as transcripts of conversations with chatbots or notes regarding litigation strategy. These fabricated citations create substantial burdens for both the courts and the opposing parties involved in legal disputes. Judges must spend extra time verifying the authenticity of every cited precedent, which can lead to significant delays in the resolution of cases. The presence of false information in legal filings complicates the pursuit of justice and places an additional strain on judicial resources. Legal professionals and court administrators are now calling for institutional responses to address this issue. The survey indicated that 91.7 percent of responding judges believe that a formal, systemic approach is necessary to combat the rise of fabricated citations. Many experts suggest that the current situation requires clear guidelines on the use of technology in legal practice. Proposed solutions include establishing legal grounds for imposing fines on parties who submit falsified documents. Approximately 61.4 percent of the surveyed judges supported the implementation of such financial penalties. Other suggested measures include requiring parties to verify the accuracy of all cited rulings before submission and mandating the disclosure of AI usage in the drafting of legal documents. Support for mandatory disclosure of AI tools remains high, with 60.6 percent of judges advocating for such a requirement. Furthermore, 55.4 percent of respondents called for potential sanctions against lawyers who rely on AI to generate inaccurate or nonexistent legal authorities. These disciplinary actions could include referrals to professional oversight bodies to ensure accountability. The findings underscore a broader global conversation about the intersection of technology and the rule of law. As AI tools become more accessible, the legal community faces the challenge of balancing innovation with the need for factual accuracy. Maintaining the reliability of legal precedents is essential for ensuring that the judicial process remains fair and transparent for all participants. Moving forward, the National Court Administration is expected to evaluate these survey results to determine the next steps for policy reform. The goal is to protect the integrity of the court system while adapting to the realities of modern digital tools. Future regulations may focus on strengthening verification processes to prevent the further spread of AI-generated misinformation in courtrooms.