Judge Admonishes Attorney Who Represents Plaintiffs in Katt Williams Case for ‘AI Hallucinations’

0
22
Federal judge schedules Rule 11 hearing after attorney files brief with fabricated AI-generated citations in Katt Williams case. Attorney blamed her daughter for the errors but faces potential sanctions and bar referral. [File Photo]

Case Intel

  • A federal judge in Georgia has scheduled a Rule 11 hearing after finding an attorney’s brief contained fabricated legal citations she attributed to her daughter’s assistance
  • Judge Ray ordered his law clerk to fact-check “all of the provisions” in the brief after noticing multiple errors, finding the “overwhelming majority” were incorrect
  • The attorney faces potential monetary sanctions, bar referral, and restrictions on future federal court filings

By SAMUEL A. LOPEZ | USA HERALD

ATLANTA — A Georgia federal judge delivered a scathing rebuke to an attorney representing four women suing comedian Katt Williams, warning she could face “serious discipline” for filing a legal brief he described as riddled with “AI hallucinations.”

During a heated hearing on Williams’ motion for summary judgment in Atlanta federal court, U.S. District Judge William M. Ray II directly confronted attorney Loletha Denise Hale about whether she used artificial intelligence to prepare her clients’ response brief—a document the judge said contained so many fabricated legal citations that he ordered his law clerk to fact-check every single one.

Signup for the USA Herald exclusive Newsletter

“After I noticed the first few [errors], I asked my law clerk to look up all of the provisions cited in it,” Judge Ray said during the proceedings. “The overwhelming majority of them were either cited incorrectly or did not stand for what [Hale] claimed they did.”

When pressed by the judge, Hale denied using AI but offered an unusual explanation: she said she had “personal issues going on” at the time and that her daughter “helped her” prepare the brief. Notably, Hale neither disclosed her daughter’s age nor confirmed whether her daughter is a licensed attorney authorized to work on the case.

“Is your signature on it?” Judge Ray asked pointedly, prompting Hale to confirm that it was indeed her signature on the document.

The judge said he found that “concerning,” not only because it presented “a possible Rule 11 issue”—referring to Federal Rule of Civil Procedure 11, which mandates that attorneys conduct reasonable inquiries into facts and law before filing papers with the court—but also because of “misrepresentations” Hale had made to a jury in another case he previously oversaw.

Judge Ray’s response signals the growing judicial intolerance for AI-generated legal work that hasn’t been properly vetted. “I’m going to notice this,” he told the parties, announcing he would soon set a date for a Rule 11 hearing where he expected there would be “a lot of testimony.”

The implications for Hale could be severe. Rule 11 hearings can result in monetary sanctions, referral to the state bar for disciplinary action, or court orders prohibiting attorneys from filing future documents without prior judicial approval. The judge’s directive that Hale “self-report to the state bar” and notify her clients of the situation underscores the gravity of the matter.

“I also think you should give notice to your client and self-report to the state bar,” Judge Ray said. “If I followed what you put in your brief, I could have denied a motion based on what you cited, which was not correct.”

Adding that Hale put her “professional reputation” at risk, Judge Ray warned that potential discipline could be “very serious” and “affect your ability to practice in federal court.”

Hale’s excuse—blaming personal issues and her daughter’s involvement—follows a troubling pattern in recent AI citation scandals. In similar cases across the country, attorneys caught filing briefs with fabricated citations have consistently pointed to personal distractions or delegated responsibility to associates, rather than taking accountability for work bearing their signature.

Rather than address the judge’s warnings directly, Hale pivoted to argue the underlying lawsuit’s merits, focusing on whether Williams was properly served with legal papers.

The underlying case stems from allegations that Williams and his entourage “jumped and threatened [the plaintiffs] at gunpoint” outside an Atlanta nightclub. The four plaintiffs—Selena Boston, Jalisa Rhodes, Lutisha Martinez, and Lanette Washington—filed their lawsuit in February 2023, but Williams wasn’t served until December 29 of that year.

Williams’ attorneys, Payden Grizzle and Gabe Banks of Banks Weaver LLC, argue this delay is fatal to the plaintiffs’ case. They contend that the statute of limitations on the plaintiffs’ claims expired on December 7, 2023, making the December 29 service too late unless the plaintiffs can demonstrate they were diligent in attempting earlier service.

“Lack of diligence with service” rendered the plaintiffs’ claims time-barred, Grizzle told the judge, pointing to the more than 10-month gap between filing and service.

Hale pushed back, arguing that serving Williams was uniquely difficult due to his tour schedule and celebrity status. She claimed it was only possible to serve him at one of his shows, but that “his security would not let anyone close enough” to complete service.

Judge Ray acknowledged serving Williams might be challenging but noted he had offered a solution in a July 2023 order: if requested, the court would authorize the U.S. Marshals Service to serve Williams “wherever he might be” and instruct “civil bodyguards” to yield to the marshals’ authority.

The judge also suggested that Williams’ performance schedule could have been discovered through depositions, to which Hale repeatedly responded, “He didn’t answer anything.”

When Judge Ray noted that formal “requests for answers” could have yielded the information, Hale claimed they made such requests but received no response. Williams’ attorneys disputed this, and Judge Ray’s review of the court docket revealed the plaintiffs had filed a “motion to determine sufficiency of responses” on July 17—suggesting some response was indeed provided. The court denied that motion days later, partly because it was untimely.

The service issue became further complicated when Grizzle noted that the plaintiffs failed to timely respond to identical requests for admission served on each plaintiff. This failure resulted in them legally admitting that Williams caused them no physical harm and that they instigated the physical altercation with his security detail.

When Judge Ray asked Hale what he should examine to determine if the plaintiffs responded timely, she claimed Williams’ former counsel, Jeremiah T. Reynolds, had granted an extension. Banks and Grizzle disputed this, though Hale produced emails she claimed proved her right. After hearing them read aloud, Judge Ray hinted he didn’t believe an extension was granted, though he said she could formally request to enter the emails into the record.

The case represents a convergence of two significant challenges facing modern litigation: the proper use of AI tools in legal practice and the persistent difficulties of serving celebrity defendants. While courts have shown increasing willingness to sanction attorneys for AI-generated errors, they’ve also recognized the legitimate challenges posed by high-profile defendants who maintain extensive security details.

The case is Boston et al. v. Williams et al., case number 1:23-cv-00752, in the U.S. District Court for the Northern District of Georgia.

The plaintiffs are represented by Loletha Denise Hale of Hale Law Firm PC. Williams is represented by Gabe Banks and Payden Grizzle of Banks Weaver LLC.