A lawsuit over a deadly 2020 explosion in Houston is drawing attention for an unexpected reason: an expert witness retained by 3M reportedly used ChatGPT to help draft key portions of a liability report in the case.

That detail is raising fresh questions about how artificial intelligence is being used in high-stakes legal disputes, according to 404 Media.

Here’s what to know

As the outlet explained, a 3M expert witness used ChatGPT, with the prompts later surfacing in discovery records, deposition materials, and court transcripts. 

Among them were 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,” 404 Media reported.

At the heart of the case is a 2020 explosion at Watson Grinding in Houston that killed three people and destroyed about 200 homes. 

The Chemical Safety and Hazard Investigation Board said the blast was caused by a “degraded and poorly crimped rubber welding hose” that leaked flammable gas, which later ignited.

Dozens of homeowners have sued Watson Grinding and 3M, alleging that 3M failed to properly maintain the facility’s gas-detection system and made other errors that contributed to the explosion. 

With hundreds of millions of dollars in liability potentially at stake, AI’s role in an expert report could become a major credibility issue.

The controversy suggests that AI use in court is no longer limited to attorneys filing chatbot-assisted briefs or citing fabricated cases. 

It is also reaching expert testimony, where judges and juries may give enormous weight to technical opinions.

More background

The expert’s prompts were discoverable, meaning opposing lawyers could review not only the final report but also the instructions that helped generate it.

Expert witnesses are typically presented as independent specialists whose conclusions are grounded in technical expertise and careful analysis. 

If parts of those conclusions were drafted or refined by a chatbot — especially one prompted to reach a specific outcome — that can complicate the appearance of the testimony’s trustworthiness.

These tools can save time, summarize dense records, and help identify patterns people might miss. 

But they can also introduce bias, factual mistakes, overconfidence, or polished language that sounds authoritative without being fully dependable.

There is also an environmental concern regarding the rapid rise of AI. 

Training and operating large AI systems require significant amounts of electricity and water, which can strain the power grid and increase utility costs for other users. 

At the same time, AI can also help optimize electricity demand, improve grid efficiency, and support cleaner energy systems such as wind and solar forecasting.

What can be done?

Courts, companies, and expert witnesses may need clearer standards for when and how AI can be used in litigation. 

Disclosure rules could help distinguish acceptable editing assistance from the use of AI to shape the substance of expert opinions in ways that could mislead juries.

Law firms and consultants can also adopt stricter internal guardrails, including requirements for human verification, rules for retaining prompts, and limits on the use of chatbots for specialized technical claims.

Will Moye, an attorney for the plaintiffs, described the incident as “really egregious,” 404 Media reported.

Get TCD’s free newsletters for easy tips, smart advice, and a chance to earn $5,000 toward home upgrades. To see more stories like this one, change your Google preferences here.