AI Risk Wire

Cases › made things up › LR-2026-0537

Attorney sanctioned for AI misuse in Hyde School lawsuit

Preliminary Why: Publisher not yet named · subject to change

Status: fix reported

Severity minor
Reported 7 May 2026
Industry legal

No source states: loss or records, developer and model, agent tool, could do, deployer.

Severity minor: test: the errors were confined to one filing by one attorney in one lawsuit and are being corrected by an amended response; no third-party system or data was stated as touched.

A lawyer for a former student suing Hyde School, a Bath, Maine boarding school, filed a response to the school's motion to dismiss that contained inaccurate and fabricated legal citations. The attorney later admitted in a new filing that the errors came from writing produced by either ChatGPT or Claude, and said she had failed to review the work of a team member before filing. The federal judge ordered her to take a course on AI in the legal field and to create procedures at her firm to prevent such errors, and imposed no fine. The judge noted that AI use is not inherently an issue but that the attorney had a duty to independently confirm the citations. The attorney accepted responsibility and expressed remorse, and she must file an amended response while the case continues.

What happened

1 · What it was doingAn attorney representing a former student suing Hyde School used AI to help draft a court filing.
2 · What went wrongThe response to the motion to dismiss improperly cited two federal cases and inaccurately quoted Maine's human trafficking law.
3 · How it was caughtIn a new filing on April 28 the attorney admitted the errors came from writing done by ChatGPT or Claude.

What was done

The judge ordered an AI legal education course and firm procedures to prevent such errors, with no fine.
Fix reported

“Court ordered an AI legal education course and firm procedures to prevent such errors”

Risks

Made-up output relied on
Primary36 incidents

Sorted by rule from the case's fields; not yet read from the sources.

Controls

Not yet read for controls. Controls that mitigate the primary risk:

Application

Grounding and fact checks · Human approval for high-impact actions · Output filtering and redaction · People told they are dealing with AI

Evidence

No source states: what it reached or did, when it happened, how many records were exposed, the direct loss, a count of the other parties affected, any court case, what set it off.

Show the full account: 6 items from the primary sources

What happened, in order

  1. April 28 Guagenty admitted in a new filing that the errors came from writing done by ChatGPT or Claude, after the court had given her 60 days to find a new Maine-based lawyer.
  2. December Guagenty addressed the issues and called the mistakes clerical errors.

Not dated in the sources:

  • Guagenty filed a response to the motion to dismiss with improper citations and inaccurate quotations.
  • The judge ordered Guagenty to take an AI legal education course and create firm procedures to prevent such errors.
  • The judge imposed no fine and Guagenty remains involved in the case.
  • Guagenty must file an amended response to the motion to dismiss, which the judge has not yet ruled on.
Show every field, with its source and grade

What happened

Earliest source we hold
2026-05-07
D inferred by us from the described behaviour, reason shown
Cause
a human check that existed was skipped or failed
C one secondary source states it
What was done
Court ordered an AI legal education course and firm procedures to prevent such errors
C one secondary source states it
What went wrong
made things up
C one secondary source states it

Impact

Severity
minor
D inferred by us from the described behaviour, reason shown
Why this severity
test: the errors were confined to one filing by one attorney in one lawsuit and are being corrected by an amended response; no third-party system or data was stated as touched.
D inferred by us from the described behaviour, reason shown

Context

Industry
legal
C one secondary source states it
Country
United States
C one secondary source states it

Notes on gaps

The sources do not state the year of the motion (last October) or of the December response, so no full dates were given. The order's date is given only as Tuesday, and the article is dated 2026-05-07, which was not used as a disclosure date.

LR-2026-0537 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON