AI News AI资讯 16h ago Updated 1h ago 更新于 1小时前 62

How courts are coping with a flood of AI-generated lawsuits 法院如何应对人工智能生成诉讼的激增

The federal judge can now tell when a lawsuit is written by a machine, and that’s both a revelation and a warning. Judge Maritza Braswell’s observation from her Colorado chambers isn’t just a quirky anecdote; it’s the frontline evidence of a seismic shift. A flood of pro se filings—lawsuits drafted by people without lawyers—is inundating federal courts, and the culprit is clear: generative AI. A new study of millions of cases shows these filings have jumped from 11% to nearly 17% of the docket i 科罗拉多州联邦治安法官 Maritza Braswell 的办公桌上,纸质文件堆成了一座小山。如今,这座小山里正悄然混入一种新的纹理:措辞流畅却偶有破绽、论点工整但引用离奇的 AI 生成诉状。她和其他法官敏锐地察觉到,那些因请不起律师或案件太小而独自步入法庭的“自助诉讼者”,数量正在激增。一份涵盖 450 万起联邦民事案件的研究给出了冰冷的数据:自诉案件比例从 2022 年的 11% 飙升至 2025 年的 16.8%,而其中 AI 参与撰写的文件数量翻了一倍不止。法官们成了第一批发现新大陆的哥伦布——这片大陆的土壤,是由算法、法条和普通人的诉讼梦共同构成的。

65
Hot 热度
70
Quality 质量
60
Impact 影响力

Analysis 深度分析

The federal judge can now tell when a lawsuit is written by a machine, and that’s both a revelation and a warning. Judge Maritza Braswell’s observation from her Colorado chambers isn’t just a quirky anecdote; it’s the frontline evidence of a seismic shift. A flood of pro se filings—lawsuits drafted by people without lawyers—is inundating federal courts, and the culprit is clear: generative AI. A new study of millions of cases shows these filings have jumped from 11% to nearly 17% of the docket in just a few years. The reason is simple: tools that were once the exclusive, expensive domain of legal professionals are now available to anyone with a prompt. But this wave isn’t bringing more justice. It’s just bringing more paper.

On its face, this is a story about access. For decades, the legal system has been a locked room for the poor and the desperate. If your case was too small for a lawyer to bother with, or your bank account too empty to hire one, you were left scribbling on scraps of paper, your arguments lost in a thicket of legal procedure you didn’t understand. Now, AI acts as a crude but effective translator, turning raw grievance into something resembling a legal pleading. Judge Braswell admits the filings are clearer, easier to parse than the "handwritten scrawls bordering on gibberish" she used to see. In that narrow sense, AI is a democratizing tool. It gives the unrepresented a voice that can at least be heard in the same language as the system.

But here’s the bitter irony: articulating a claim better is not the same as winning it. The study finds no improvement in outcomes for these AI-assisted litigants. This exposes the core fallacy in the "AI will democratize law" narrative. Justice isn’t just about having your complaint typed in a professional format. It’s about strategy, evidence, procedural nuance, and the ability to navigate a system designed by lawyers, for lawyers. AI can mimic the form of a legal argument but cannot provide the judgment of a lawyer who knows which hill to die on, which motion to file, and when a case is truly hopeless. It’s giving people a louder megaphone, but it’s still shouting into a void. The system’s barriers are strategic and interpretive, not just linguistic. AI doesn’t lower those; it just makes the shouting more eloquent.

This creates a peculiar new burden for the judiciary. Judges are now forced into a dual role: arbiter of the law and de facto AI-content moderator. They’re spotting hallucinated case citations and fabricated quotes woven into filings. They have to read charitably, but how do you charitably interpret a legal authority that doesn’t exist? It adds a layer of forensic work to an already overwhelmed docket. Worse, it threatens to degrade the quality of legal precedent. If judges begin to see made-up statutes and phantom rulings cited repeatedly by pro se litigants, the very foundation of case law—a shared universe of existing decisions—starts to erode. The system relies on a shared factual baseline; AI, in its current form, is actively contaminating it.

The most profound question, though, is the one the article touches on at the end: responsibility. We are entering a legal twilight zone where a chatbot performs the functional role of a lawyer—giving legal advice, drafting documents—while being shielded by every possible disclaimer from the duties that attach to that role. A human lawyer has a fiduciary duty to their client; they can be sued for malpractice, disbarred for incompetence. What duty does an AI tool have? None. Its outputs are protected by terms of service buried in legalese. It’s a system that wants all the societal utility of a professional service while accepting none of the accountability. When the AI hallucinates a winning case that doesn't exist, and a desperate person files a lawsuit based on that fiction, who bears the cost of that wasted time, that crushed hope? The user, for being naive enough to trust a machine? The platform that profitably served them the lie? Or a society that allows this duty-free practice of pseudo-law?

Lawmakers are finally starting to circle this problem, but they’re years behind the technology’s deployment. They’re debating who should be liable for bad AI legal advice while the advice is already flooding their courts. The entire framework of legal ethics was built for human relationships of trust and expertise. We’re now grafting a tool onto that framework that is fundamentally incompatible with its assumptions. An AI is not a junior partner learning the ropes; it’s a stochastic parrot with a law degree it didn’t earn.

So what are we left with? A story of profound ambivalence. AI is undeniably expanding who can engage with the legal system. That’s a net good in a society that claims to believe in equal justice under law. But it may simultaneously be cheapening the very notion of what constitutes a valid legal claim, while creating an accountability vacuum at the center of a new, automated legal industry. The uptick in pro se filings isn’t a triumph of technology over bureaucracy. It’s a canary in the coal mine, signaling a future where the justice system is deluged not with more justice, but with more procedurally compliant, intellectually hollow, and ethically orphaned paperwork. The judge can understand the filings better. But what she’s understanding is increasingly a hollow mimicry of justice, and that might be more dangerous than the gibberish it replaced.

科罗拉多州联邦治安法官 Maritza Braswell 的办公桌上,纸质文件堆成了一座小山。如今,这座小山里正悄然混入一种新的纹理:措辞流畅却偶有破绽、论点工整但引用离奇的 AI 生成诉状。她和其他法官敏锐地察觉到,那些因请不起律师或案件太小而独自步入法庭的“自助诉讼者”,数量正在激增。一份涵盖 450 万起联邦民事案件的研究给出了冰冷的数据:自诉案件比例从 2022 年的 11% 飙升至 2025 年的 16.8%,而其中 AI 参与撰写的文件数量翻了一倍不止。法官们成了第一批发现新大陆的哥伦布——这片大陆的土壤,是由算法、法条和普通人的诉讼梦共同构成的。

AI 在这里扮演的角色,远非简单的“工具”二字可以概括。它是一个充满矛盾的“新租户”,猛地闯入司法系统这间老房子。一方面,它以近乎民主化的姿态,填平了法律文书的专业鸿沟。过去那些如同天书、让法官耗尽耐心辨认的手写“狂草”,正在被 AI 梳理成相对清晰的论点。Braswell 法官承认,她如今处理 AI 辅助的自诉状甚至更快。这听起来像个双赢故事:司法效率提升,当事人声音得以被更好地“听清”。但法官那句“能理解得更好一点,我或许就能更好地帮助他”的潜台词,恰恰揭示了残酷的真相:AI 帮你把话说清楚了,但没帮你把道理说对,更没帮你赢得官司。胜诉率的纹丝不动,像一盆冷水,泼在了“AI 赋能司法公正”的浪漫幻想之上。

更危险的裂缝,出现在责任的真空地带。当人类律师向法庭呈上一份文件,其背后是整个执业资格体系、行业伦理规范和法律责任的重压。一个错误的引证,可能导致其执照被吊销。而现在,一个聊天机器人在几分钟内炮制出看似专业、实则充满虚构判例和捏造引语的诉状,然后呢?当这份包含“幻觉”的文件被提交,当它导致诉讼延误甚至败诉,责任该由谁来扛?是点击鼠标的、不懂法律的用户?是开发模型却声明“仅供参考”的科技公司?还是那个本就超负荷运转的司法系统?美国各地的立法者们开始抓耳挠腮,因为这个问题暴露了现有法律框架的深层窘迫:我们正在用 20 世纪的责任体系,去管辖一个 21 世纪的、由代码生成的新主体。

这绝非杞人忧天。AI 正在将法庭变成一个巨大的“试衣间”。无数人涌入,抱着 AI 生成的“完美诉状”试一试运气,希望这个神奇工具能帮他们赢得赔偿或主张权利。科技公司则在一旁微笑,将“提高司法可及性”写入漂亮的年度报告,却对“可及性”不等于“胜诉可能性”这一关键区别闭口不谈。他们的商业模式建立在规模化服务上,而诉讼案件的激增正是完美的燃料。他们提供的是一把万能钥匙的幻觉,却把锁匠(人类律师)的专业性、法院的甄别成本、以及当事人真实的机会成本,都悄悄转嫁给了整个系统。Braswell 法官这样的“技术娴熟者”或许能一眼识破 AI 的破绽,但更多资源紧张的法院呢?

所以,当我们讨论 AI 对法律的冲击时,焦点不应再仅仅停留在“AI 能否替代律师”这种过时的二元对立上。真正的议题更为深刻且紧迫:一个不承担律师职业责任、不拥有法律实体资格、却能大规模影响司法进程的“行为者”,我们该如何为其划定边界?是要求所有 AI 生成的法律文件必须像律师签名一样,由某个可追溯的实体担责?还是建立强制性、高标准的“法律 AI”认证,确保其输出符合事实核查?亦或,最务实的做法是,明确在法律文书中标识 AI 参与,让法官和对方律师获得知情权?

司法系统正面临一场静默的“殖民”。AI 不是来摧毁法庭的,它是来改变其生态的。它让声音更容易被听见,但并未让正义的天平更容易倾斜。它带来的便利是真实的,但随之转移的责任风险也是巨大的。当 Braswell 法官们不得不花更多精力去甄别 AI 的“幻觉”时,这本身就是一种系统效率的隐性损耗。科技公司不能一边享受着用户增长和数据红利,一边用“用户自行负责”的免责条款置身事外。是时候建立新的规则了——不是为了扼杀创新,而是为了确保当算法的脚步踏入神圣的法庭时,它留下的不是更多难以收拾的混乱,而是对程序正义基本敬畏。否则,我们迎来的将不是一个更公正的司法未来,而是一个充斥着更流畅废话、更精致谎言的数字泥潭。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

法律AI 法律AI 监管 监管 伦理 伦理
Share: 分享到: