How to Write a Legal Brief With AI Assistance
AI can meaningfully speed up brief writing — structuring arguments, drafting first-pass language, checking citation formatting — but it can’t replace the legal judgment a brief actually depends on. Here’s how to use it well without the risks that have gotten attorneys sanctioned.
Where AI genuinely helps
Outlining and structuring: feeding an AI tool your key facts, the applicable standard of review, and your core arguments, and asking it to propose a logical structure, is a legitimate and fast way to get past a blank page. Drafting boilerplate sections (procedural history, standard of review recitations) that follow a predictable format is another strong use case, since these sections are formulaic by nature. AI is also useful for tightening prose — cutting redundant language, improving clarity and flow — once the substantive argument is already yours.
Where it’s genuinely risky
Case citations are the single biggest danger. General-purpose AI models (ChatGPT, Claude, Gemini) can generate citations to cases that sound completely plausible and don’t exist, or that exist but don’t say what the AI claims they say. This isn’t a rare glitch — it’s a well-documented pattern, and it has led to real court sanctions against attorneys who filed AI-drafted briefs without independently verifying every citation. Never cite a case in a filed brief that you have not personally pulled up and read.
The workflow that actually works
Draft your argument’s substance and legal theory yourself, or with a legal-specific research tool (CoCounsel, Westlaw’s AI features) that retrieves from real case databases rather than generating from a general language model’s memory. Use a general AI tool for structure and prose polish on top of that verified substance. Then independently verify every single citation, quote, and factual claim in the final draft before filing — treat this step as mandatory, not optional, regardless of which tool you used.
Legal-specific tools vs. general AI for briefs
CoCounsel and similar legal AI platforms are meaningfully safer for the research and citation-finding portion of brief writing because they retrieve from actual legal databases rather than generating citations from a general model’s training data. That said, even these tools require verification — “safer” doesn’t mean “verified for you,” and treating any AI output as citation-ready without a human check is the exact behavior that has led to sanctions.
Court and jurisdiction rules on AI disclosure
A growing number of courts have adopted local rules requiring disclosure when AI was used in drafting a filing, or certification that citations have been verified. These rules vary significantly by jurisdiction and change fairly often, so checking your specific court’s current standing order or local rules on AI use before filing is essential — don’t assume the rule in one court applies in another.
The bottom line
AI is a legitimate productivity tool for brief writing when used for structure, drafting speed, and prose quality — and a serious professional liability risk when treated as a substitute for legal research and citation verification. The attorneys who’ve gotten in trouble weren’t using AI per se; they were filing its output without checking it.