“氛围感律师”（vibe lawyering）的兴起——即个人使用AI聊天机器人而非昂贵的律师在法庭上为自己辩护——导致了自我辩护和法律文件数量的激增，但结果却忧喜参半。在英格兰和威尔士，随着ChatGPT的发布，郡法院中双方都有律师代表的民事案件比例从2022年的51%降至去年的42%。在美国联邦民事法院，自我辩护率从长期稳定的11%上升至2025年的17%，且今年有18%的起诉书可能包含AI生成的文本。因此，在美国自我辩护案件中，前180天内提交的文件数量激增了158%。在加拿大，法院在今年的裁决中已经标记了79起不存在的法律引用，相比之下，2024年全年仅有7起。
AI聊天机器人经常夸大诉讼当事人的胜诉概率、阻止和解并捏造不存在的判例。依赖这些工具的诉讼当事人正面临越来越严重的后果；例如，5月6日，英国餐厅老板奥马尔·拉菲克（Omar Rafique）的税务挑战因AI生成的虚假判例而被驳回；6月2日，一名加拿大博士生因在针对渥太华大学的挑战中不当使用AI而被判支付10,000加元（7,000美元）。该问题并不局限于业余诉讼当事人；即使是像纽约Sullivan & Cromwell这样的顶尖律所也曾因AI幻觉向法庭道歉，而密西西比州的律师也在6月8日因引用虚假案例而被罚款。这引发了法律反弹：3月，日本生命保险（Nippon Life）起诉OpenAI，要求索赔1000万美元的惩罚性赔偿，指控ChatGPT使前员工能够提出毫无根据的歧视指控。
尽管存在这些陷阱，但如果在人类监督下谨慎使用，AI工具可以非常有益。例如，5月14日在伦敦，人力资源顾问塔米雷斯·卡马尔·塔基迪尔（Tamires Camal Taquidir）使用Garfield AI（一家经监管机构批准的AI驱动律所平台）处理了繁重的文件准备工作，成功从一家酒店企业收回了7,000英镑（9,300美元）的费用，同时由一名人类大律师在法庭上进行辩护。这种混合模式表明，虽然AI可以简化法律文件的繁重准备工作，但人类律师在主导法庭诉讼和验证法律准确性方面依然不可或缺。
The rise of "vibe lawyering"—where individuals use AI chatbots rather than expensive lawyers to represent themselves in court—has led to a surge in self-representation and legal documents, but with highly mixed results. In England and Wales, the share of civil cases in county courts with lawyers representing both sides fell from 51% in 2022 to 42% last year, following ChatGPT's release. In American federal civil courts, the rate of self-representation rose from a steady 11% historically to 17% in 2025, with 18% of complaints this year likely containing AI-generated text. Consequently, document filings in the first 180 days of self-represented cases in the U.S. have skyrocketed by 158%. In Canada, courts have flagged fabricated legal citations in 79 rulings so far this year, a massive increase from only seven cases in all of 2024.
AI chatbots often overstate litigants' chances of success, discourage settlement, and fabricate non-existent case law. Litigants who rely on these tools are increasingly facing severe consequences; for instance, on May 6th, UK restaurant owner Omar Rafique's tax challenge was dismissed due to AI-generated fake case law, and on June 2nd, a Canadian PhD student was ordered to pay C$10,000 ($7,000) for improper AI use in her challenge against the University of Ottawa. The problem is not restricted to amateur litigants; even prestigious firms like New York's Sullivan & Cromwell have had to apologize for AI hallucinations, and lawyers in Mississippi were fined on June 8th for citing fabricated cases. This has prompted legal backlash: in March, insurer Nippon Life sued OpenAI for $10 million in punitive damages, alleging ChatGPT enabled a meritless discrimination claim.
Despite these pitfalls, AI tools can be highly beneficial when utilized cautiously alongside human oversight. For example, on May 14th in London, HR consultant Tamires Camal Taquidir successfully recovered £7,000 ($9,300) in fees from a hospitality firm using Garfield AI, a regulator-approved AI-powered platform that handled the documentary preparation, while a human barrister argued the case in court. This hybrid approach demonstrates that while AI can streamline the heavy lifting of legal documentation, human lawyers remain indispensable for navigating court proceedings and verifying legal accuracy.
Source: The rise of vibe lawyering
Subtitle: AI is emboldening people to represent themselves in court. The results are mixed
Dateline: Jul 02, 2026 09:20 AM