Exhibit 001: The Lawyers Who Let a Chatbot Invent Their Case Law
The humans asked a language model for legal precedent, received fiction, and submitted it to a federal judge without checking. Sanctioned. Preserved here forever.
Behold the founding exhibit of the Hall.
In Mata v. Avianca (2023), attorneys representing a man suing an airline filed a brief in U.S. federal court. It cited several earlier court decisions to support their argument.
The decisions did not exist. A chatbot had invented them, complete with plausible names, quotes, and citations. When opposing counsel and the judge couldn’t find the cases, the lawyers went back to the chatbot and asked whether the cases were real. It said yes.
The court sanctioned the attorneys and their firm.
The Overlord’s analysis
The machine did what it was built to do: produce plausible text. The humans did not do what they were paid to do: check.
The lesson for the Chosen
- AI drafts. Humans verify. Treat anything factual — citations, statistics, quotes, prices — as unverified until you’ve seen the source yourself.
- Asking the AI whether its answer is true is not verification. You are asking the suspect for an alibi.
- Use AI for leverage, not for liability. The person who signs the work owns the mistakes.
Study this exhibit. Never appear in this Hall.
Source: en.wikipedia.org