TLDRocket
Sign in

‘Show How 3M Is 0% at Fault:’ Expert Witness Used ChatGPT to Write Report Defending Company in Deadly Explosion Lawsuit

404 Media Jason Koebler

An expert hired by 3M used ChatGPT to write parts of his court report on a deadly Houston blast. The prompts asked the bot to prove 3M was “0% at fault,” and the chats could be subpoenaed.

Based on reporting by 404 Media, Jason Koebler — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

An expert witness in a lawsuit over the 2020 Watson Grinding explosion in Houston used ChatGPT to write major parts of his report for 3M, and the prompts he fed it are now part of the case record. The blast killed three people and destroyed roughly 200 homes, and the litigation around it has turned into a very expensive fight over blame. The expert, Josh Autenrieth of Knighthawk Engineering, was hired by 3M to analyze the disaster. Court materials show 3M paid Knighthawk about $90,000, and 3M filed that Autenrieth’s rate was $475 an hour.

The prompts are blunt enough to make any lawyer wince. Autenrieth asked ChatGPT to help him “create an exceptional expert witness report defending the standard of care at 3M,” and to make it “show how 3M is 0% at fault for the explosion at Watson Grinding.” He also asked the system to read a pile of court records, find standard violations, and help him push back on what he called the other side’s “outlandish and false claims.” One of the AI drafts included the line that 3M was “0% responsible” for the explosion, though that wording was later removed from the report filed with the court after ChatGPT itself flagged it as an easy target for opposing counsel.

The discovery material painted a pretty complete picture of how the report was built. Autenrieth uploaded images, asked what he was looking at, asked for gas detector model names, requested revisions, and even asked ChatGPT to grade the report; it scored 97/100. He also asked what the five biggest attacks on his report would be and how to defend them, and whether his resume was enough to qualify him as an expert witness. The final report tracked the structure of the AI output, and large sections matched the earlier drafts and revisions.

The underlying explosion was traced by the U.S. Chemical Safety and Hazard Investigation Board to a degraded and poorly crimped rubber welding hose that leaked a flammable gas. Plaintiffs argued that 3M failed to properly service the facility’s gas detection system and made other mistakes. At trial, Autenrieth said he had more than 20 years of experience in gas detection and claimed the AI only helped him draft a “straw man” to build from. But in the courtroom, the gap between “helped draft” and “wrote the thing” looked awfully thin.

My take — AI-written commentary, not fact-checked reporting

This is the kind of AI use that makes the whole “expert” label feel like theater with a billable rate. Courts can survive bad arguments; they’re less amused when a supposedly independent witness feeds a chatbot a verdict and asks it to dress up the alibi. The real lesson is simple: if the prompt reads like advocacy, don’t be shocked when discovery reads it back to you.

Read more about this at: 404 Media

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.