Skip to content
Chimera readability score 65 out of 100, Academic reading level.

The Volokh Conspiracy
Mostly law professors | Sometimes contrarian | Often libertarian | Always independent
Florida Politician (Anthony F. Sabatini) Faulted for AI Hallucinations in Briefs
"It is rare to see the kind of blatant and repeated misconduct that Sabatini [acting as plaintifs' lawyer] has committed in this case."
From Akerlund v. Atlas Air, Inc., decided yesterday by the Eleventh Circuit, Judge Britt Grant, joined by Judge Robin Rosenbaum and Embry Kidd:
A group of plaintiffs, employees in the commercial aviation business, personally reject their companies' pandemic-era policies on masks, testing, and vaccination. The district court dismissed the third amended complaint for lack of personal jurisdiction and failure to state a claim, and we affirm.
The claims in this case are remarkably weak, at least as pleaded. We are more candid than usual in this assessment because the plaintiffs' counsel Anthony F. Sabatini has not been candid with us. Sabatini filed multiple briefs replete with fake and hallucinated citations. Even after being warned. "Always a bad idea." Chief Justice John G. Roberts, Jr., 2023 Year-End Report on the Federal Judiciary, at 6 (2023). By outsourcing his legal work to an AI algorithm, Sabatini violated his ethical duties to both his clients and this Court….
Though the filings in this case were substandard in a variety of ways [I excerpt some of the substantive analysis below -EV], we have saved the worst for last. The plaintiffs' counsel Anthony F. Sabatini's filings are riddled with citations to nonexistent, "hallucinated" cases. His opening brief relies on at least eight such cases, including one purportedly decided by this Court.
After the defendants identified this problem, Sabatini acknowledged in his (untimely) proposed reply brief that those citations were "erroneous or unverifiable," and sought to withdraw his reliance on eight listed cases. At this point, things go from bad to worse: the eight cases Sabatini "withdrew" did not match a single one of the eight hallucinated cases in his opening brief. And not only were they not the right cases—all eight were also hallucinated.
We are far from the first court to see lawyers uncritically rely on artificial intelligence software and submit briefs citing nonexistent cases. The typical response when this kind of violation is identified is some version of an apology from the attorney, often with sanctions to follow. It is rare to see the kind of blatant and repeated misconduct that Sabatini has committed in this case.
The first rule of our profession is that a lawyer "shall provide competent representation to a client"—"competent" in the sense that it requires "legal knowledge, skill, thoroughness and preparation." It goes without saying that completely outsourcing one's legal work to artificial intelligence software is not competent. Doing so is a dramatic violation of the client's interests.
But it does not stop there, because lawyers are also officers of the Court. We expect that when lawyers submit briefs, they give us their best view of what the law is, and how that law supports their clients. We read those briefs carefully—not just because they are the product of counsel's time, effort, and skill, but because they help us as we try to reach the right answer. All that is lost if a lawyer decides that it is not worthwhile to do the work to persuade us—if he is willing to sign his name on whatever a machine churns out without so much as checking it for accuracy. We cannot do our job the same way unless lawyers do theirs.
Whatever the merits of artificial intelligence, it is no substitute for actual intelligence. Any "use of AI requires caution and humility." When lawyers rely on AI tools, there is no way to get around the obligation to verify that the software got it right, and that what it got wrong will not end up in court filings.
{AI algorithms are notorious for producing "outputs that echo users' opinions and beliefs, even when those views are incorrect." This tendency presents a particular danger in legal filings, as one of a lawyer's chief duties is to give his clients a clear-eyed view of whether the law says what the client wants it to say.}
By signing his name on briefs filled with hallucinated citations—not once but twice—Sabatini violated his professional obligations, both to his clients and to this Court. If he thinks these claims are not worth the effort, he should either tell his clients they fall short or advise them to hire another lawyer—not file obviously insufficient pleadings and briefs generated by AI tools. In a separate order, this Court, through the Chief Judge, will refer the matter to the Committee on Lawyer Qualifications and Conduct.
Orlando Sentinel (Annie Martin) reports that Sabatini is an elected County Commissioner for Lake County, Florida, and a former state representative:
The 37-year-old has made headlines for years for his controversial and often insulting social media posts and was warned by Republican leaders in Tallahassee to tone down his remarks even before he even attended his first legislative session in 2019.
In the past, he has tweeted a picture of an AR-15 targeted at George Floyd protesters, called for ending gay marriage and, when he ran unsuccessfully for the U.S. House of Representatives four years ago, posted on X, "I'm running for Congress to imprison as many Democrats as possible."
And a few excerpts related to the substance:
Atlas Air is a commercial airline, and Flight Services International hires contractors to staff Atlas Air's flights. During the Covid-19 pandemic, both companies required employees to vaccinate against Covid-19, unless they obtained a religious or medical exemption—in which case, they had to undergo periodic testing and wear a mask on the job.
The plaintiffs say they object on religious grounds to what they perceive as "a dangerous social and medical experiment." They explain that, among other things, their "conscience prohibits them from being inoculated with any experimental foreign substance," and that their religious rights were violated as a result of "the Biden Administration's goal of achieving universal vaccination and to unlawfully acquire [their] personal, genetic information." And in their view, the companies' accommodations for religious objectors to the vaccination requirement were unreasonable: monthly testing imposed "substantial burdens," and masks were "functionally useless" "symbols" that accomplished "nothing more than political advertising." Some plaintiffs say that they "succumbed to the pressure" and took the vaccine; others begrudgingly wore masks and tested. The complaint does not allege that anyone lost their job, but does allege that some plaintiffs were assigned to less desirable, lower-paying flights.
The plaintiffs assert that their employers' Covid-19 protocols led to a hostile work environment under Title VII; a Federal Food, Drug, and Cosmetic Act violation; a federal constitutional deprivation; a tortious invasion of privacy; a negligent disclosure of private medical information; and an infliction of emotional distress. The district court dismissed all claims against Flight Services International and some claims against Atlas Air for lack of personal jurisdiction, and the remaining claims against Atlas Air for failure to state a claim….
Exercising personal jurisdiction over Flight Services International in this lawsuit would violate due process because neither general nor specific jurisdiction is available in Florida for this company….
The plaintiffs first allege that Atlas Air created a work environment hostile to their religious beliefs, in violation of Title VII. They say that Atlas Air's "repeated attempts to coerce" them into getting vaccinated constituted "unwelcomed harassment."
To prevail on this claim, "the plaintiff is required to prove that the defendant had a discriminatory intent or motive." But the plaintiffs do not allege any facts indicating that Atlas Air intended to discriminate against their religious beliefs. In fact, the allegations suggest the opposite. Under its policy, Atlas Air exempted from its vaccine requirement employees who professed a sincere religious objection to the Covid-19 vaccine. Instead, they had to wear a mask at work and test once a month.
The plaintiffs now seem to assert that these accommodations are also hostile to their religion. The problem with this argument is that there is nothing to back it up in the pleadings. The plaintiffs do not allege any religious objections to masks or tests—only political and logistical ones. Masks, they say, are "functionally useless" "political symbols" that "serve no other purpose than identifying 'dissident' employees." And the plaintiffs offer no real problem with testing, other than a generalized complaint that it entails "substantial burdens." What those burdens are, and whether they are religious in nature, the plaintiffs do not say. Nor do they allege that employees who received religious exemptions were treated any worse than those who received medical exemptions.
Whatever one makes of the plaintiffs' dislike of masks and tests, they have alleged no facts signaling animosity from Atlas Air toward their religious beliefs. The allegation that Atlas Air harbored a discriminatory motive is not only conclusory, but "wildly implausible." …
The plaintiffs also bring claims under the Federal Food, Drug, and Cosmetic Act, alleging that Atlas Air unlawfully required its employees to take medical products authorized for emergency use. But they cannot sue to enforce such claims; only the federal government can bring enforcement actions under that statute….
The plaintiffs also allege several state tort claims. The first is that Atlas Air tortiously invaded their privacy by disclosing private medical information—namely, vaccination status—to company administrators in charge of enforcing Covid-19 protocols. "In Florida, except in cases of physical invasion, the tort of invasion of privacy must be accompanied by publication to the public in general or to a large number of persons." And publicity "requires that a matter be made public, by communicating it to the public at large, or to so many persons that the matter must be regarded as substantially certain to become one of public knowledge." The plaintiffs do not allege sufficient facts on this score….
We also reject the plaintiffs' intentional infliction of emotional distress claims. Those would require showing that Atlas Air intentionally or recklessly caused "severe emotional distress" through "extreme and outrageous conduct." The challenged conduct must "go beyond all possible bounds of decency, and to be regarded as atrocious, and utterly intolerable in a civilized community."
Like many other employers at the time, Atlas Air required certain precautions designed to prevent the spread of Covid-19. But the fact that the plaintiffs disagree with the steps Atlas Air took does not put the airline's actions beyond all possible bounds of decency. The district court correctly dismissed the plaintiffs' intentional infliction of emotional distress claims.

Facts Only

* Plaintiffs' counsel Anthony F. Sabatini filed briefs with citations to nonexistent or "hallucinated" cases.
* Sabatini acknowledged in a proposed reply brief that those citations were erroneous or unverifiable and sought to withdraw reliance on eight listed cases.
* The eight cases Sabatini withdrew did not match the eight hallucinated cases cited in his opening brief.
* The court found that outsourcing legal work to AI violates the professional duty of providing competent representation.
* Lawyers are expected to provide the court with their best view of the law, necessitating accuracy in filings.
* Plaintiffs alleged claims against Atlas Air and Flight Services International based on COVID-19 protocols, religious objections, privacy, and hostile work environment.
* The district court dismissed claims against some defendants due to lack of personal jurisdiction or failure to state a claim.
* In the context of the Atlas Air case, plaintiffs failed to allege facts indicating discriminatory intent for Title VII claims regarding religious accommodations.
* Plaintiffs failed to allege sufficient facts regarding publication necessary for an invasion of privacy claim under Florida law.
* Intentional infliction of emotional distress claims were dismissed by the district court due to lack of evidence of "extreme and outrageous conduct."

Executive Summary

Plaintiffs' counsel Anthony F. Sabatini filed multiple legal documents containing citations to cases that the court found to be nonexistent or hallucinated, including one purportedly decided by the court itself. When defendants identified this issue, Sabatini acknowledged in a proposed reply brief that those citations were erroneous or unverifiable and sought to withdraw reliance on eight listed cases. However, the eight cases Sabatini withdrew did not match the eight initially hallucinated cases. The court found that outsourcing legal work to AI violated professional duties, as competence requires legal knowledge, skill, thoroughness, and preparation. Furthermore, relying on AI in legal filings violates the obligation for lawyers to present their best view of the law to the court, necessitating verification. The matter was referred to the Committee on Lawyer Qualifications and Conduct.
The underlying dispute involves plaintiffs who objected to employer COVID-19 protocols, including mask mandates and testing, based on religious objections and privacy concerns. Claims asserted included hostile work environment under Title VII, Federal Food, Drug, and Cosmetic Act violations, constitutional deprivation, invasion of privacy regarding medical information, and infliction of emotional distress. The court dismissed claims against some defendants due to lack of personal jurisdiction or failure to state a claim. Analysis of the specific claims indicated that plaintiffs failed to allege necessary facts, such as discriminatory intent, for certain tort claims, and the conduct did not meet the high bar for intentional infliction of emotional distress.

Full Take

The situation reveals a critical friction point between technological capability, professional ethics, and judicial integrity. The misconduct demonstrated by Sabatini highlights a danger where procedural shortcuts, enabled by tools like AI, can undermine the very foundation of legal representation—the duty of competence. The pattern observed is not merely an isolated instance of error but a systemic risk: when the mechanism for generating persuasive legal argument (AI) bypasses the necessary human verification process, the resulting output can introduce falsehoods directly into the judicial record. This challenges the premise that legal advocacy functions as a careful presentation of established law; instead, it risks becoming a machine-generated echo chamber that prioritizes fluency over fidelity to reality.
The underlying narrative involves specific, context-heavy claims regarding pandemic protocols and religious accommodation, which are themselves being navigated through an ethically compromised procedural lens. The plaintiffs’ assertions regarding mask efficacy or the scope of employer liability were ultimately judged on their factual basis rather than the method of presentation. This suggests a tension between external political or social grievances (the protesters, vaccination mandates) and the internal, verifiable requirements for legal argumentation. The implication is that while AI can generate sophisticated text, the responsibility for ensuring that text reflects actual legal reality—and not mere statistical plausibility—remains squarely with the human actor, whether lawyer or litigant.
What if the systemic cost of outsourcing critical judgment to algorithms shifts the burden of proof entirely onto the client? If an attorney operates in a space where high-stakes factual and ethical verification is mandatory, allowing reliance on unverified outputs creates a scenario where error becomes institutionalized noise, eroding the capacity for genuine legal discovery. The lack of required scrutiny acts as a form of semantic camouflage, making it easier to obscure flawed reasoning under the guise of sophisticated output. How does the concept of "competent representation" translate when the very means by which representation is generated are fundamentally opaque? What recourse remains when the tools designed for efficiency actively sabotage the adversarial process that demands transparent truth-seeking?

Sentinel — Human

Confidence

The text reads like structured legal commentary synthesizing a court decision with a critique of the role of AI in legal practice, grounded in specific facts from the case.

Florida Politician (Anthony F. Sabatini) Faulted for AI Hallucinations in Briefs — Arc Codex