"AI Said So?"
A Growing Trend in Global Courtrooms
"Legal Profession Polarized by Ability to Verify AI"

Recently, the British current affairs weekly The Economist ran an in-depth report on the side effects of artificial intelligence (AI), bringing the new term "Vibe Lawyering" to the forefront in the global legal community. Literally translated, Vibe Lawyering means "atmosphere advocacy." It refers to the practice where attorneys or parties involved in a case base their arguments solely on emotional appeals, intuition, or the mood of the courtroom instead of citing established legal logic, statutes, or precedents. Without proper legal backing, they may insist, "This doesn't make sense from a common-sense perspective," or "Anyone can see this is unfair," in an unreasonable manner. The Economist reported that as AI technologies become more widespread, there has been a surge in "pro se" lawsuits—cases without legal representation—leading to the growing prevalence of this issue.


Courts in Korea are also not immune to side effects arising from the use of generative AI. Judges on the front lines have already recognized that a considerable number of attorneys are using AI to draft legal documents. Distinctive features can be noticed in the documents themselves, such as writing style and structure.


In fact, there have recently been cases where briefs submitted to the court began with "Gemini Response." A presiding judge at a district court noted that during a trial, he encountered an unfamiliar legal principle in a written brief for the first time. When he pointed this out, the representative was flustered and later resubmitted a corrected version. There was strong suspicion that the principle had been fabricated by AI, but as no evidence could be found, the judge was unable even to issue a warning about AI use. Another presiding judge shared that when he searched for the case number of a precedent cited in a pro se litigant's brief, he discovered that such a precedent did not exist in the court’s system.


According to statistics provided by the National Court Administration, as of 2025, cases in which either the plaintiff or defendant did not appoint legal counsel accounted for 46.23% of civil panel cases, 78.78% of single-judge civil cases, and 97.43% of small claims cases. When combining all three types, 89.24% of all civil cases are handled as pro se litigation.


According to the legal tech platform Courtready, Canadian courts identified a total of 79 judgments involving fake precedents so far in 2026. In the UK, when ChatGPT was launched in 2022, 51% of civil cases had both parties represented by lawyers, but this figure dropped to 42% last year. Joint research by MIT and USC found that the proportion of pro se litigation in U.S. federal civil courts stayed around 11% for a long period, but surged to 17% in 2025. The researchers analyzed that approximately 18% of complaints filed in 2026 contained AI-generated text. Recently, the major New York law firm Sullivan & Cromwell officially apologized after citing AI-fabricated fake precedents in their court documents.


Joo Woo-hyuk, head of the International Legal Affairs Team at Dongin Law Group, explained, "Vibe Lawyering has begun to appear not only in the legal communities of English-speaking countries but also recently in Korea. If an attorney relies solely on creating an atmosphere without grounding arguments in legal principles, this could result in disciplinary action or liability for breach of the duty of care." He added, "There are precedents where, if there was no direct attorney involvement in using AI to assess legal risks, the attorney-client privilege may be lost." In the United States District Court for the Southern District of New York, it was ruled that a document outlining a litigation strategy independently drafted by a client using public AI (Claude), without the guidance of an attorney, is not protected by attorney-client privilege (ACP) or work-product doctrine (United States v. Heppner).


It is also pointed out that in the Korean legal community, there is an urgent need for countermeasures as the number of pro se litigants who do not verify AI-generated answers and persist with their arguments in court continues to rise. Lee Sewon (Judicial Research and Training Institute, 37th class), managing attorney at Law Office Seohwadam, stated, "The legal profession is becoming increasingly polarized in terms of capability, depending on whether attorneys have the ability to verify AI-generated answers."


A presiding judge who requested anonymity commented, "Currently, legal practitioners lack the ability to control and filter AI errors. Only those who can control AI and distinguish fake legal principles should be granted legal qualifications."




Jisoo Kim, Law Times Reporter


※This article is based on content supplied by Law Times.

This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing