Attorney for Maine Client Faces Sanctions for AI-Driven Errors in Court Filing
A federal judge in Maine sanctioned an attorney representing a former Hyde School student after AI-assisted drafting introduced fabricated and inaccurate legal citations into court filings, highlighting growing scrutiny of generative tools in legal practice.
Attorney for Maine Client Faces Sanctions for AI-Driven Errors in Court Filing
A court rebuke over AI-generated legal mistakes
A federal judge in Maine has sanctioned attorney Kelly Guagenty, who represents a former student suing Hyde School in Bath, after court filings in the case included inaccurate and fabricated legal citations linked to the use of generative AI tools. The underlying lawsuit accuses the boarding school and related defendants of misconduct including forced labor, human trafficking, and negligence, but the immediate focus shifted to the reliability of the plaintiff’s legal briefing and the responsibilities of counsel when using AI in litigation.
Judge Stacey D. Neumann concluded that the filings contained serious citation problems and that the errors were significant enough to warrant sanctions, though not monetary penalties. Guagenty was allowed to remain on the case, but the ruling made clear that technology does not lessen a lawyer’s obligation to verify every legal authority presented to a court.
What went wrong in the filings
The disputed filing was submitted in response to a motion seeking dismissal of the lawsuit. In that response, Guagenty cited federal cases and legal authorities that did not accurately support the arguments being made, and some quotations and references could not be verified. A later attempt to fix the problems through an errata filing introduced additional inaccuracies rather than resolving the court’s concerns.
As the issue unfolded, Guagenty eventually acknowledged that the errors stemmed from an AI-assisted drafting process involving either ChatGPT or Claude. She said she had failed to conduct a careful, line-by-line review of the work before it was filed. That admission came only after the court issued a show-cause order asking why sanctions should not be imposed.
The judge treated that sequence as important. The problem was not simply that AI had been used, but that the resulting authorities were not independently checked before being presented as valid legal support.
“Although AI can be a useful aid in research and drafting, its use does not diminish an attorney’s nondelegable duties.”
The sanctions and the judge’s reasoning
Rather than imposing a fine, the court opted for non-monetary sanctions aimed at deterrence and professional correction. Guagenty was ordered to complete continuing legal education related to generative AI, create clearer internal procedures at her law firm to prevent similar failures, provide the sanctions order to her client, and certify to the court that she had complied.
The court also struck the flawed opposition filing from the docket and required the plaintiff to submit an amended response to the motion to dismiss. That means the larger case against Hyde School is still alive, but the plaintiff’s side must now refile its arguments without the defective citations.
Judge Neumann’s order emphasized that AI is not inherently improper in legal practice. What matters, in the court’s view, is human verification, candor, and professional judgment. The ruling described the risks of unverified AI output as broader than a single bad filing: opposing counsel must spend time uncovering errors, courts must divert attention to policing invented authorities, and public confidence in the legal system can be damaged when false citations appear in official pleadings.
The people and institutions involved
The case centers on a lawsuit brought by Jessica Fuller, a former student, against Hyde School, a well-known boarding school in Bath, Maine. Guagenty, a Massachusetts attorney, had initially been able to appear in the case with the support of Maine-based attorney John Steed, who later withdrew after the citation issues surfaced. Guagenty later secured new local counsel and told the court she had taken steps to improve oversight within her firm.
The judge also noted mitigating factors. Guagenty accepted responsibility, expressed remorse, and did not have a prior disciplinary record. Those factors appear to have helped her avoid a harsher penalty such as monetary sanctions or removal from the case.
Guagenty said the court’s decision was fair under the circumstances and indicated that she would comply with the order while continuing to represent her client.
Why the case matters for technology and AI
The dispute reflects a larger shift in the legal profession as generative AI becomes more common in drafting, research, and document preparation. Courts increasingly recognize that these systems can produce text that looks authoritative while containing invented cases, inaccurate quotations, or unsupported legal claims. In high-stakes settings like litigation, that makes unsupervised or poorly supervised AI use especially risky.
For the technology sector, the ruling is another example of how adoption is colliding with accountability. The key lesson is not that legal professionals cannot use AI, but that AI outputs must be treated as unverified drafts, not reliable authority. The burden remains on the attorney whose name is on the filing.
That lesson extends beyond law firms. Any profession using generative tools for specialized work, including education, healthcare, compliance, and public administration, faces the same basic challenge: automation can accelerate drafting, but it cannot replace responsibility for accuracy.
Broader relevance beyond Maine
There is no direct connection to California’s Central Valley in the dispute itself, but the warning is broadly relevant to courts, law firms, schools, and public institutions across the region. As organizations in places such as Fresno, Bakersfield, Modesto, and Merced adopt AI tools for research and writing, the case underscores a practical point that applies everywhere: efficiency gains mean little if the underlying material is wrong.
For Central Valley institutions, the significance lies in governance. The ruling reinforces the need for internal review rules, staff training, and disclosure practices whenever generative systems are used in important filings or formal documents. In that sense, the case is part of a wider national test of how traditional professional standards will be enforced in an AI-assisted workplace.
Central Valley AI is produced by the CVAI Newsdesk team and developed by Kaweah Tech, a regional firm that builds, deploys, and integrates AI solutions for businesses across California's Central Valley.