A law firm used ChatGPT to justify its fees - and the judge wasn't convinced

What happened

Cuddy Law Firm used ChatGPT as part of its justification for a requested legal fee - arguing for more than $113,000, with ChatGPT cited as a supporting "cross-check."

Judge Paul Engelmayer found the use of ChatGPT unpersuasive and ultimately awarded around $53,050 - roughly half what was requested. There was no sanction like in Mata v. Avianca; the problem was that AI output was used as an argument in a context where it had no real evidentiary weight.

The business problem

The issue wasn't that the AI produced something false - it was using AI in a role it wasn't suited for in the workflow.

What could have helped

• Classify what AI is (and isn't) an appropriate input for
• Don't let AI output substitute for evidence or professional judgment
• Put human-review requirements around anything client- or court-facing
• Train staff on where AI adds credibility and where it undermines it

Where this points