AI slopSomeone handed you AI slop. Now it is your problem.

The client's AI slop

A client runs their contract through a general purpose AI chat, say ChatGPT, and sends you the result asking for a fast yes.

You open it expecting a sign-off and find terms that drift between sections and a clause that does not do what they think.

You quote review time, and they ask why a finished-looking document needs billable hours.

On the next matter, the same kind of draft arrives.

Opposing counsel's AI slop

Opposing counsel, or a pro se party, files something the chatbot clearly helped write.

Smooth in places, generic in others, confident about law it never quite cites.

You cannot act on a hunch, so you read the whole filing and check which citations exist before arguing what they say.

You catch the fabricated case this time. The next filing arrives the same way.

The client's AI slop, reviewed not redrafted

You open the draft knowing where these break: definitions that shift between sections, a clause that reads clean and does the wrong thing.

You clear the citations first, separately from the substance.

Cite Check checks identified legal citations against available legal sources and surfaces citation and treatment information for attorney review. A successful match does not establish that an authority supports a particular proposition; attorneys should review the underlying authority before relying on it.

Opposing counsel's AI slop, checked not guessed

You run the citations separately from the substance, before you read for meaning.

Cite Check checks identified legal citations against available legal sources and surfaces citation and treatment information for attorney review. A successful match does not establish that an authority supports a particular proposition; attorneys should review the underlying authority before relying on it.