Company case

Morgan & Morgan: eight made-up cases from the firm’s own AI tool, signed by three lawyers

A lawyer asked his firm’s own AI tool to add case law to a court filing. Eight of the nine cases it cited did not exist. Two other lawyers signed without reading. The firm had trained staff, and it then added a warning inside the tool.

Outcome Went wrong

Something failed in work where AI was used, and there is a public record of it.

In the guide What fails

At a glance

Company
Morgan & MorganLaw firm
Size
Enterprise (1,000+ employees) The firm’s website says it has more than 1,100 attorneys.
Industry
Professional services
Where and when
United States (federal court in Wyoming). January – February 2025.
Who produced the figures
Court or tribunal recordThe facts and quotations come from the court’s published order on sanctions. The headcount is the firm’s own.

What happened

The firm acted for a family suing a retailer and a manufacturer. On 22 January 2025 its lawyers filed a set of motions before trial. One lawyer had written the draft. He then uploaded it to the firm’s in-house AI tool and asked it to “add more case law”.

The other side could not find the cases and said so. The court asked the lawyers to explain.

The numbers

  • Nine cases cited. Eight did not exist.
  • Three signatures on the filing. Only one of the three lawyers had worked on it.
  • $3,000 fine for the lawyer who wrote it. He was also removed from the case.
  • $1,000 fine each for the two lawyers who signed without reading.
  • No penalty for the firm.

What the court found

  • A tool built by the firm made up the cases. This was not a public chatbot. It was the firm’s own system.
  • Signing means checking. The two other lawyers said they had not read the filing. The court said that made no difference. A signature confirms that the signer has read the document and checked the law in it. That duty cannot be handed to someone else.
  • The errors were visible. The court noted that the fake cases were not in the usual citation format. A reader who looked would have seen it.
  • Openness reduced the penalty. The lawyers withdrew the filing, were “honest and forthcoming about the use of AI”, paid the other side’s fees and showed what they had changed. The court named the drafter’s “honesty and candor” as a reason for a lighter penalty.

What the firm had done, and what it changed

The firm showed the court that it had trained employees not to use the tool this way. After the court’s order it added a step inside the tool. Before using it, a person must accept a notice that “users must independently verify” anything the AI produces.

The court said it would probably have ordered similar measures itself, so it did not penalise the firm.

What the case shows

  • Your own AI tool can invent things too. Building a tool in-house does not remove the need to check its output.
  • Training did not stop one person. A rule that lives in a training session can be forgotten. The firm moved the rule into the tool, at the point of use.
  • A second signature is not a review. Two people approved a document neither had read.

What this says about the practices

Practices the company applied

Practices that were missing

What the company changed afterwards

Sources

  1. Wadsworth v. Walmart Inc., No. 2:23-cv-118-KHR. Order on Sanctions and Other Disciplinary Action, 24 February 2025. US District Court, District of Wyoming. storage.courtlistener.com
  2. Morgan & Morgan. About us (the firm’s own description of its size). forthepeople.com