Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
The reported use of ChatGPT by an expert retained by 3M in litigation over Houston’s Watson Grinding explosion illustrates a central AI-governance risk: a tool can help produce expert work, but it cannot supply independence, verify evidence, or take responsibility for conclusions. Reporting says the expert’s prompts and related AI material surfaced in discovery and drew scrutiny at trial. That makes the case a practical lesson in verification, records, disclosure, and expert accountability—not proof that AI use automatically invalidates expert evidence.
What happened in the 3M ChatGPT case?
In an August 17, 2026 report, 404 Media’s Jason Koebler said engineering expert Josh Autenrieth, associated with Knighthawk Engineering and retained by 3M, used ChatGPT to help prepare significant portions of an expert report. The report describes discovery exposing prompts and related AI material, followed by questioning about how the work was prepared.
404 Media reproduced prompts attributed to Autenrieth, including requests to “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.” These are reported prompt excerpts, not findings or language from a court ruling. CBS News, in a September 18, 2026 report by Alyssa Spady, said the AI use drew scrutiny at trial and cited legal experts’ concerns about reliability. That reported commentary is not a judicial determination that the report or testimony was unreliable.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteThe underlying explosion dispute
The litigation concerns the January 2020 Watson Grinding explosion in Houston. According to the U.S. Chemical Safety and Hazard Investigation Board description quoted by 404 Media, a degraded, poorly crimped rubber welding hose leaked flammable gas before the explosion. The report said three people died and homes were damaged or destroyed. Plaintiffs alleged that 3M’s work on a gas-detection system contributed to the incident; that allegation should not be confused with an established finding of fault.
#1 Best Overall
Why do the reported prompts raise an independence concern?
An expert’s value depends on analysis that can be explained as independent and grounded in evidence. A request for a report defending a client’s standard of care could be consistent with a retained expert’s assignment to assess a defense position. But a prompt asking a model to show that the client is “0% at fault” frames the task around a predetermined result. That creates an appearance problem: readers may reasonably ask whether the analysis followed the evidence or was shaped to reach the requested conclusion. This is a governance inference from the reported prompts, not a court holding about Autenrieth’s work.
AI-generated language can make a report sound coherent without showing that its reasoning is sound. The professional signing or presenting the work remains responsible for understanding and explaining its sources, assumptions, calculations, and conclusions. Using an AI tool does not, by itself, establish that a claim is true or that an expert independently reached it.
What does the case reveal about AI governance?
Preserve expert control of the analysis
AI can assist with drafting or organizing material, but the expert should decide what evidence matters and reach conclusions through a process they can defend. A useful review question is: could the expert explain why each material conclusion follows from the evidence without relying on the model’s authority?
Check claims and citations, not just prose
Fluent writing is not evidence of accuracy. Every material factual assertion, quotation, citation, calculation, and description of a source needs a human check against the underlying material. 404 Media’s account described a “Citation Overlay” and discovery of AI interactions; the practical lesson is to make source traceability part of review, rather than treating a polished document as verified.
Rank #3
Plan for records and disclosure before using AI
Prompts, inputs, outputs, and drafts may become relevant in litigation. In this case, 404 Media reported that prompts and AI-related documents emerged in discovery. That does not establish that every prompt is always discoverable: relevance, privilege, protective orders, and procedural circumstances can matter. Still, organizations and experts should agree in advance on what to retain, how to protect confidential information, who reviews AI-assisted material, and how the method will be disclosed when appropriate.
What should organizations and experts put in place?
The following are practical governance measures suggested by the case, not requirements shown to have been imposed by a court:
- Define permitted use. Specify whether AI may be used for research, outlining, drafting, summarization, or other tasks, and identify any prohibited uses.
- Keep the expert responsible. Require the expert to review and own the final analysis, including every material assertion and conclusion.
- Make verification auditable. Preserve the source trail for claims and citations, and document checks of calculations and quotations.
- Set a retention and confidentiality process. Decide how prompts, outputs, drafts, and inputs will be handled, including how sensitive or protected information will be treated.
- Agree on review and disclosure. Clarify who examines AI-assisted work and how the expert and client will address questions about the process.
These controls address distinct risks: independence concerns whether the expert directs the reasoning; verification concerns whether the work is supported; records practices affect traceability; and confidentiality and litigation obligations shape what can safely be entered into a tool and how material is handled afterward.
What does this case not establish?
The reporting supports a case-specific account of AI use, discovery, and trial scrutiny. It does not establish a general rule that using ChatGPT invalidates expert testimony, that every AI prompt must be disclosed, or that a court found the expert’s report unreliable. The available reporting also does not establish the final judgment, any post-trial rulings, or the appeal status. A reported jury award alone is not enough to determine the case’s final procedural outcome.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
