ChatGPT prompts used by expert witness to write report defending 3M in explosion at Watson Grinding presented in court, here’s the ‘important lesson’ it gives
3M hired an expert witness to prepare a report about a 2020 Houston explosion that killed three people and destroyed roughly 200 homes. The witness reportedly used ChatGPT to write significant portions of his expert report, with the use of AI coming to light after prompts he used while preparing the report were revealed during court proceedings.The prompts showed that Josh Autenrieth of Knighthawk Engineering asked ChatGPT to help him “create an exceptional expert witness report defending the standard of care at 3M” and to “show how 3M is 0% at fault for the explosion at Watson Grinding.” During testimony, Autenrieth acknowledged using ChatGPT and said the submitted report was “90 to 85% ChatGPT.”
ChatGPT prompts used for the 3M expert report
The case concerns the January 2020 explosion at Watson Grinding, a Houston manufacturing facility. The US Chemical Safety and Hazard Investigation Board said a degraded, poorly crimped rubber welding hose caused the explosion by leaking flammable gas.Dozens of homeowners sued 3M and Watson Grinding over the explosion. Plaintiffs alleged that 3M failed to properly service the facility’s gas detection system and made other errors that contributed to the incident.As part of the litigation, 3M hired Autenrieth to prepare an expert report addressing the company’s role in the explosion. During discovery, plaintiffs’ attorney Will Moye found a five-page document called “Citation Overlay” that appeared to have been generated using AI.Moye asked 3M’s lawyers for the prompts Autenrieth had used. The deposition was paused for three hours while the material was collected, and Moye received about 350 pages of ChatGPT conversations.Autenrieth had told ChatGPT that he was being retained by 3M to defend the company and needed to prepare an expert report concerning its standard of care. He uploaded hundreds of court records and instructed the chatbot to review them and support 3M’s position.ChatGPT initially generated a roughly 30-page report containing the statement, “From a technical and standard-of-care standpoint, 3M is 0% responsible for the January 24, 2020 explosion.” That wording was later removed after Autenrieth asked ChatGPT to assess the report from the perspective of opposing counsel.
Expert witness used ChatGPT to revise and assess report
The ChatGPT conversations showed that Autenrieth repeatedly uploaded versions of the report and asked the chatbot to revise them. He also asked it to grade the report and identify arguments the opposing side could challenge.At one point, ChatGPT gave the report a score of 97 out of 100. Autenrieth also asked it to identify the five main issues in the report that the opposing side could attack and how he could respond.In another conversation, he uploaded an image of a gas detector and asked ChatGPT, “What am I looking at?” He also asked about model names of industrial gas detectors. The gas detection system was central to the plaintiffs’ allegations against 3M.Autenrieth also asked ChatGPT whether his resume was sufficient for him to serve as an expert witness and whether his lack of previous experience as an expert witness could be raised during the case.The report eventually submitted to the court followed the structure of ChatGPT’s initial output, while substantial portions were identical to the chatbot’s original response and subsequent revisions.Autenrieth acknowledged using ChatGPT but maintained that his professional experience remained the basis for his opinions.“I’ve got a body of work and 20-plus years of experience in the industry,” Autenrieth said.He described the AI’s role by saying, “My opinions were put in there, and AI helped me to draft a straw man to build off of.” He also said, “I put information and opinions in up front before it ever generated […] if the output wasn’t of my opinion or what I agreed to, I did alter it.”Autenrieth said he also used ChatGPT while “sitting in traffic” to “jog my memory” about different manufacturers of gas detectors.
ChatGPT conversations became part of court discovery
The disclosure of the prompts also highlighted how AI conversations used while preparing litigation materials can become relevant during discovery or depositions. Moye said the case showed why lawyers need to consider how their expert witnesses use AI when preparing reports.“This expert relied on AI not as an assistive device, but exclusively relied on ChatGPT to form his opinions and write his report,” Moye told 404 Media. “He acknowledged [at trial] the prompts he put in were biased toward 3M to help 3M win the case […] it’s really egregious.”Moye said 3M later tried to have Autenrieth disqualified from the trial. After learning about the extent of his AI use, Moye instead called the opposing expert as a witness so the jury could hear about the report.The jury ultimately awarded more than $61 million to the plaintiffs and assigned 30% of the responsibility to 3M and 70% to Watson Grinding. Earlier decisions had also awarded $118 million and $38 million to victims of the explosion.The case is among several instances in which AI-generated material has emerged in legal proceedings. In France, AI was used to make an argument about rainwater patterns in a flooding dispute, while in Canada, a judge gave no weight to an AI-generated opinion concerning plants and a bug infestation.The Watson Grinding case shows that prompts and AI conversations can form part of the record when an expert’s use of an AI system becomes relevant to litigation. For lawyers and expert witnesses, the case also demonstrates that using an AI chatbot while preparing court testimony does not necessarily keep the underlying prompts outside the discovery process.
