Expert witness used ChatGPT to write liability report in $61M lawsuit

The buried detail is the prompt itself, which explicitly instructed ChatGPT to reach a predetermined conclusion. That is not hallucination or careless use; it is deliberate outcome engineering, which shifts the culpability analysis from the tool toward the professional who deployed it and the firm that submitted the report.

This is largely disconnected from recent activity in the Modelwire archive, as we have no prior coverage of AI in legal proceedings or expert witness accountability. The story belongs to a broader pattern, documented outside our archive, of professionals submitting AI-generated work product without disclosure, most visibly in the Mata v. Avianca case in 2023 where attorneys cited fabricated case law. What is new here is the explicit bias instruction, which moves the conversation from accidental misuse to something courts may treat as fraud on the tribunal rather than mere negligence.

Watch whether the presiding court in this case issues a formal sanction referencing the AI-generated origin of the report, and whether any state bar association opens a disciplinary inquiry within the next six months. Either action would set a precedent that forces disclosure rules into litigation practice far faster than any voluntary industry standard would.