Vibe coding is a software development method in which users use plain-language prompts to get artificial intelligence to write code. The idea has since spread into law. “Vibe lawyering” refers to a growing trend of attorneys using chatbots to research case law and prepare legal documents.
The problem starts when people trust that text without checking its sources. Court filings have already exposed the consequences, including fabricated cases and made-up quotes.
From Code to Courtrooms
Lawyers are expensive and court procedures can be confusing. People without legal training can ask a chatbot at 2 AM without getting billed by the hour. For someone facing eviction or a contract dispute, an AI-generated brief offers a fighting chance on a budget.
Research by Anand Shah and Joshua Levy reveals activity has grown among people representing themselves in U.S. federal civil cases. The study found self-represented litigants accounted for 16.8% of cases in fiscal year 2025, compared with a long-term average of 11%.
This does not mean that every self-represented litigant uses AI, or that AI caused every increase in filings. However, chatbots do lower the barrier to producing legal documents. A person who once struggled to understand a court form can ask a chatbot to generate their argument in seconds.
The Hallucination Hazard
A large language model predicts which words are statistically likely to come next, based on patterns in its training data, rather than looking anything up in the moment.
That process typically produces something reasonable. Occasionally, however, it produces something entirely invented yet delivered with the same fluent tone as a correct answer. This is an AI hallucination. In the legal field, a discipline where factual accuracy carries serious consequences, failure matters.
Fabricated Cases and Fake Precedents
Several court cases have exposed what happens when users accept AI-generated legal research without checking it. Lawyers and self-represented litigants alike have submitted briefs built around case names that don’t exist or citations that, when fact-checked, say nothing of the kind.
In 2023, a judge fined two lawyers and their law firm after they submitted erroneous ChatGPT-generated citations in a court filing. After the judge called the fabricated legal cases ChatGPT cited into question, the lawyers doubled down and didn’t immediately admit to using AI. As a result, they faced sanctions and fines.
Other incidents followed. From January to June 2025, Reuters reported at least three instances involving fabricated citations in cases including Kohls v. Ellison, Coomer v. Lindell and Shahid v. Esaam.
The pattern repeats because the failure is invisible until someone checks. A fabricated case citation reads exactly like a real one, down to the formatting and conventions — nothing about its appearance flags it as fiction.
The Dangers of Deepfakes and AI Impersonation
Vibe lawyering is one symptom of a broader reliability problem in generative AI. The same models that invent case law can just as easily produce a convincing voice clone of a company’s chief financial officer or a phishing email tailored to a specific employee’s job title.
Attackers now rely on generative AI tools to create realistic messages and voice clones, automating a form of personalized deception that once required real research and effort. AI helped generate 80% of phishing emails in 2025.
Why AI Can’t Replace Lawyers
AI has useful roles in legal work. It can summarize documents, organize information, generate first drafts and help people identify questions that need further research. The user must still check the model’s output against authoritative sources.
Courts and legal professionals have used forms of AI for decades. By the 1990s, AI-powered voice recognition tools had entered dictation software and call centers. Modern speech-to-text systems can save lawyers time by producing transcripts or draft text, while users check the result against the original recording.
AI handles repetitive tasks while a lawyer retains responsibility for decisions and client communication. The American Bar Association’s Formal Opinion 512 also advises lawyers to understand generative AI’s capabilities and limitations, review its output, protect client information and maintain appropriate supervision.
How AI Flaws Create Permanent Legal Debt
IT teams can patch a software bug overnight, but a legal error filed with a court stays on the record. A fabricated citation may undermine an entire filing, or a careless prompt may expose confidential information. Each mistake creates work that lawyers, clients and courts must spend time correcting.
Breaching Confidentiality
Confidentiality creates a major concern for anyone using public systems. Legal professionals routinely handle information involving clients, financial records, trade secrets, litigation strategy and other sensitive material. Pasting that information into a general-purpose chatbot creates questions about where the data goes and how the provider handles it.
It is not enough to instruct the model not to store user data or avoid revealing private information. The exposure happens as soon as a user sends a request containing sensitive information. This data may appear in debugging tools, analytics events, request logs, test fixtures or retry payloads. It can leak through any system that touches the prompt.
Facing Sanctions and Dismissals
Courts have started imposing real consequences when AI-generated material enters legal proceedings without proper verification. A public database tracking AI-related court errors recorded more than 1,600 cases involving hallucinated legal material by June 2026.
Sanctions can include fines, reprimands, suspended proceedings and other penalties. Fact-checking and investigating citations and correcting filings also require additional time, which incurs additional costs.
The problem extends beyond a single incorrect citation. When a filing contains repeated fabrications, a judge may question the reliability of the rest of the submission. AI-generated text can call the credibility of the entire filing into question, affecting the underlying claim.
Navigating the Future of AI in Law
Vibe lawyering shows how easily AI can enter professional work. Chatbots can help with research and drafting, but users still need to protect confidential information and take responsibility for the final work. In court, confident AI-generated text carries little value when its facts and citations cannot withstand scrutiny.