Verbatim · The explainer series

Why legal AI invents citations.

Fabricated cases, misquoted holdings, statutes made to say something they don’t — “hallucinations” aren’t random. They’re driven by specific, recognizable conditions of a legal task. This series names each cause, shows how it shows up in a brief, and explains what removes it before you file.

What Verbatim does
The map

Three kinds of cause.

The conditions that spike a brief’s hallucination rate aren’t a single problem. They fall into three groups, and each has a different fix.

Demand-side causes come from the ask — the position you need, how much authority you demand, the shape of the template, and the voice you write the prompt in.

Supply-side causes come from the law and LLM training itself — where real authority is thin, where it has changed since the model was trained, and where the actual holding has been buried under mountains of media talking points, political rhetoric, and even non-lawyer mythology.

Process-side causes come from how the draft is generated — long chains of reasoning, long documents that build on themselves, and whether anything checks the citations before they’re used.

Begin

Verify the brief before you file the brief.

Verbatim reads a finished brief and reports, for every authority it cites, whether the cite is real and whether the quoted language actually appears at the pin cite — so a fabrication surfaces on your screen, not in a show-cause order. Bring a brief and we’ll walk you through the report.