Court Obtains 350 Pages of ChatGPT Logs in 3M Explosion Case
Updated
Updated · Futurism · Aug 18
Court Obtains 350 Pages of ChatGPT Logs in 3M Explosion Case
3 articles · Updated · Futurism · Aug 18
Summary
350 pages of ChatGPT transcripts were turned over after a judge ordered discovery into how 3M expert Josh Autenrieth prepared his defense report in the Houston explosion lawsuit.
The logs showed Autenrieth asked ChatGPT to "create an exceptional expert witness report" proving 3M was "0 percent at fault," uploaded hundreds of case records, and even asked the bot to identify a gas detector central to the case.
Will Moye, a plaintiffs' lawyer, said the expert used AI not as assistance but to form his opinions and draft the report, calling the conduct egregious.
The case already ended with a $61 million plaintiff verdict that found 3M 30% responsible for the January 2020 Watson Grinding blast, which killed three people and damaged about 200 homes.
The disclosure underscores a wider legal risk: ChatGPT prompts and transcripts can be subpoenaed in discovery, exposing how experts and lawyers generate work product.
When an AI warns an expert witness about bias, could those hidden chat logs destroy a multimillion-dollar defense strategy?
If generative AI drafts a legal expert's opinion, who is truly testifying on the stand—the human or the machine?
$61.5 Million Verdict Against 3M Exposes Dangers of AI-Generated Expert Testimony in Court
Overview
The 2020 Watson Grinding explosion in Houston led to devastating losses and a wave of lawsuits against 3M. To defend itself, 3M relied on expert Josh Autenrieth, who secretly used ChatGPT to write most of his technical report. Plaintiffs’ attorneys uncovered this through suspicious language and a deep review of his drafting history, revealing hundreds of pages of AI prompts. When Autenrieth admitted this under oath, the jury lost trust in 3M’s defense, resulting in a $61.5 million verdict against the company. This scandal has forced courts and legal teams to demand transparency about AI use, setting new standards for expert credibility and discovery.