Client and Lawyer Both Responsible for Attorney Fees in AI Hallucination Case

From Pauliah v. Univ. of Miss. Medical Center, decided last Tuesday by Judge Carlton Reeves (S.D. Miss.):

Courts across the country have dealt with the rising misuse of generative artificial intelligence to prepare court filings. Those cases have largely, if not entirely, dealt with citations to non-existent legal authority or the attribution of quotes to cases that do not contain the quoted material—produced as a result of what has come to be termed “AI hallucinations.” This case is different [from other AI hallucination cases], as it appears that AI was used not to hallucinate the law, but to hallucinate the facts.

The declaration at issue contained multiple fabricated quotations, presented to the Court along with manufactured citations to deposition transcripts, as if they came from sworn testimony. The declaration also grossly mischaracterized testimony and other facts in the record. See Docket No. 141 at 4-6 (listing four outright fabricated quotations and other misrepresentations made to the Court). This declaration was filed in opposition to a motion for summary judgment. Counsel expressly used some of these fabricated “facts” to argue to the Court that this case contained genuine issues in factual dispute. Manufacturing “facts,” then presenting them to the Court as genuine, threatens to corrupt the Court’s analysis and undermine the integrity of the judicial process at the summary judgment stage.

The crux of the Court’s ruling on a motion for summary judgment is determining the existence or non-existence of genuine issues of material fact. To make this determination, the Court relies on submissions from the parties. A party’s submission of fabricated “facts,” hinders the issuance of court rulings.

The lies and mischaracterizations submitted in this case substantially slowed the judicial process, as it required opposing counsel, then the Court, to dedicate significant resources to first determine whether the “factual material” before the Court was even true, prior to considering any legal implications that may flow from these “facts.” It also precipitated what would have otherwise been unnecessary filings: Defendants’ motion to strike; Plaintiff’s response; Defendants’ reply in support of the same. It also altered Defendants’ reply in support of summary judgment.

This Court’s concern is not merely for the misuse of generative AI and the inexcusable submission of fabricated “facts,” slowing this process and wasting resources; the Court is also deeply concerned by the refusal of either Plaintiff or his former counsel to accept responsibility for creating and submitting these fabrications. They have had multiple opportunities to do so.

Plaintiff and his former counsel were first confronted with these falsehoods in Defendants’ motion to strike, which identified many of the fabrications and misstatements contained in the declaration. When confronted with these fabricated quotations, however, they failed to take any responsibility. See, e.g., Docket No. 131 at 8 (explaining the false quotation, cited as Howard Dep. 33:5–34:3, by stating “Dr. Howard’s testimony discusses the Plaintiff not receiving a laptop” and entirely avoiding the portion of the false quotation purporting that Plaintiff “did not receive adequate departmental support”); id. (claiming, that “Plaintiff’s affidavit comports with Dr. Howard’s page 26 deposition testimony” despite the fabricated quotation — “No accreditation failure was ever cited against Dr. Pauliah” — appearing nowhere in Dr. Howard’s testimony).

After this Court found that the declaration was filed in bad faith, the Court ordered Defendants to serve on Plaintiff submissions Defendants previously provided to the Court for in camera review, showing the hourly rates and hours spent by counsel and staff who performed work on behalf of Defendants in response to the filing of Plaintiff’s declaration. Text-Order dated Nov. 4, 2025. The Court then provided Plaintiff and his former counsel three weeks after receipt to review those submissions and file objections to the same…. In his response, Dr. Pauliah doubled down and rejected all responsibility, asking the Court to instead sanction his former counsel. Neither objected to the hourly rates identified by Defendants.

Following this exchange, Plaintiff and his counsel were again afforded an opportunity to take responsibility for filing falsehoods with the Court. During the 56(h) hearing, Dr. Pauliah and his former counsel, Mr. Begley, disagreed over who drafted which portions of the bad faith declaration. Largely, they blamed each other. Mr. Begley claimed he drafted about half and Dr. Pauliah drafted about half. By contrast, Dr. Pauliah claimed he drafted merely two to three paragraphs in the more than 40-page, single-spaced declaration. Dr. Pauliah’s assertion is preposterous. {Dr. Pauliah has filed emails in the record which appear to show him taking an active role in the response to Defendants’ motion to strike. See Docket No. 162-14 at 21 (email from Dr. Pauliah to Mr. Begley dated August 25, 2025—the day before Plaintiff’s response to the motion to strike was filed—appearing to encourage Mr. Begley to avoid taking any responsibility for the fabrications, stating “There is one additional correction needed on page 9, lines 1 and 2. Kindly remove the text highlighted in green: ‘Rather, it appears to be entirely fabricated.‘ The corrected sentence should read: Response: Dr. Howard’s deposition testimony comports with paragraphs 139 and 147.”) (emphasis in original). In addition, at the end of the hearing, Dr. Pauliah admitted to using AI—offering as his only excuse, that everyone does.} Regardless, neither is blameless. Mr. Begley acknowledged at the hearing that he had a responsibility to review whatever Dr. Pauliah prepared, and Dr. Pauliah admitted, after some questioning, that he used generative AI to draft at least a portion of his declaration. He also admitted he did not review his declaration prior to signing it.

Mr. Begley’s actions are unacceptable. The Court understands that Mr. Begley and other lawyers may be unfamiliar with artificial intelligence, the hallucinations it has shown itself prone to, or how their clients may use or misuse it. That said, lawyers are obligated to maintain the standards imposed on them as officers of the court, including forthrightness and candor. Those obligations predate artificial intelligence by centuries.

How Mr. Begley did not recognize or uncover Dr. Pauliah’s citations as completely fabricated deposition testimony is hard to fathom, as Mr. Begley not only attended the depositions, he took them, and he obtained copies of the transcripts. Simply turning to the purported source of any of these quotations—each was accompanied by a manufactured citation to a deposition transcript, ostensibly to lend the appearance of authenticity—would have revealed them as fake. Further, once Defendants filed their motion to strike, Mr. Begley should have discovered the fabrications and, upon his discovery, been honest with the Court. Mr. Begley did not meet the demands of his professional obligations here.

Dr. Pauliah’s behavior is also unacceptable. He admitted to signing a declaration, under the penalty of perjury, without verifying—or even attempting to verify, it seems—the truth of its contents. He has yet to accept responsibility for the role he played in submitting these fabrications to the Court and then denying that they were fabrications, once confronted. Mr. Begley’s obligation to review his client’s declaration does not absolve Dr. Pauliah. Whether he recognizes it or not, Dr. Pauliah may not draft his own declaration with disregard for the veracity of its contents. Just above his signature, he declared: “under penalty of perjury that the foregoing is true and correct.” It was not.

The Court finds that sanctions are appropriate in this case. Neither Mr. Begley nor Dr. Pauliah objected to the Defendants’ rates for counsel or staff. Using those rates, the Court finds that 40.4 hours of work, totaling $8,570, are directly related to identifying and responding to the fabrications contained in the offending declaration. This amount includes the 10.9 hours, totaling $2,140.50, to which Mr. Begley did not object. Upon taking into consideration Plaintiff’s and his former counsel’s economic circumstances, however, the Court finds that $5,000 is an appropriate amount to levy against Mr. Begley [$4,000] and Dr. Pauliah [$1,000]. {At the 56(h) hearing, Mr. Begley represented that he is no longer in private practice. In his response and objections to Defendants’ fee submission, Mr. Begley further represented that he “is a salaried employee of Jackson city government, which is his chief source of income, and [he] does not have professional malpractice insurance to cover this matter.” He also represented that Dr. Pauliah “has not been employed since he left UMMC, as evidenced by his tax returns that he provided the defense in discovery.”}

Here’s a statement from Dr. Pauliah about the case (sent in response to my query to him and Mr. Begley about this):

The court’s order raises profound concerns because it constructs a novel ‘AI hallucination facts’ theory out of thin air — a theory with no evidentiary basis, no grounding in the record, and no connection to established legal practice. Justice is truth, and that truth has no premise in this order or in the way these proceedings unfolded. The full evidentiary record is already docketed, yet the ruling does not engage with the corrective filings, clarifications, and exhibits I submitted to address the manifest injustice that had occurred. Those materials were before the court but were overlooked and replaced with speculation. When a decision departs from the record and relies on unsupported assumptions, the integrity of the process is compromised.

Because the existing record already contains the evidence necessary to resolve these issues, I believe a careful review of the docketed materials will make the discrepancies clear — including the absence of any factual basis for the court’s ‘AI hallucination facts’ theory. These structural and procedural errors go to the core of judicial fairness and due‑process requirements. They also reflect a broader national concern: courts invoking artificial‑intelligence concepts without evidence, notice, or procedural safeguards. I welcome continued public attention to the serious implications this decision presents.

And here’s a statement from Mr. Begley:

I don’t have any comments on the Court’s analysis.

Judge Reeves acted within his discretion to make the Rule 56(h) bad faith ruling, strike the declaration of the plaintiff, and assess attorney’s fees against me.

I will not be appealing this ruling.

I delivered a $4,000 check to the opposing law firm the day after the Court entered the order.

In accordance with the Order of the Court, I plan to participate in this CLE seminar: Ethical Considerations When Using AI to Draft Legal Documents.

No event in my professional life of 40 years has been more damaging or distressing than this sanction.

Chelsea C. Lewis and Malissa Wilson (Forman Watkins & Krutz, LLP) represent defendants.