Hulvat v. Gumina: Illinois Appellate Court Grants Sanctions for AI-Hallucinated Brief in Post-Divorce Dispute | Advocate Prakhar

Case at a Glance

Background

Robert Hulvat’s marriage to Jennifer Hulvat was dissolved in 2004. Post-dissolution proceedings in 2017 generated a dispute over the terms of an agreed order concerning college expense contributions. Through an admitted error, Jennifer’s attorney (Gumina) submitted a prior version of the agreed order to the court, which entered the erroneous version. The parties subsequently agreed in writing to correct the error, and the court eventually vacated the erroneous order in April 2023.

Robert, not content with the eventual correction, filed a stand-alone civil lawsuit in 2022 against Jennifer, Gumina, and Gumina’s law firm alleging civil conspiracy and “unauthorised filing” — essentially, that the submission of the wrong order was a fraud. The circuit court dismissed the complaint, finding that Gumina’s conduct was protected by absolute litigation privilege — statements and filings made in the course of judicial proceedings are absolutely privileged, regardless of their correctness or the parties’ intent.

Robert appealed pro se. His appellate brief cited nonexistent cases and fabricated holdings, triggering a rule to show cause from the Court of Appeals and a motion for sanctions by the defense.

The AI Issue

The Illinois Court of Appeals found that Robert’s appellate brief violated Illinois Supreme Court policy and rules on AI use, citing citations to nonexistent authorities and fictitious holdings. The court simultaneously resolved the show cause regarding AI misconduct and the merits of the appeal — finding the appeal substantively meritless (absolute litigation privilege barred the claims) and the brief procedurally deficient (AI hallucinations). Both deficiencies contributed to the court granting defense sanctions.

What the Court Decided

• Dismissal of civil conspiracy and unauthorised filing counts affirmed: absolute litigation privilege barred claims arising from Gumina’s filing of the agreed order in judicial proceedings, even if the wrong version was filed.
• Forgery count (against Gumina alone) also barred by absolute litigation privilege.
• Rule to show cause for AI-hallucinated brief resolved: court found the brief violated Illinois Supreme Court policy and the appellate rules governing citation accuracy.
• Defendants’ motion for sanctions granted: attorneys’ fees incurred in defending against the improper use of AI in the appellate brief are awarded as a sanction.
• The combination of meritless substantive claims and AI-hallucinated legal authorities produced an appeal that was both factually and legally defective.

“Plaintiff’s misuse of artificial intelligence in his appellate brief violates Illinois Supreme Court policy and rules, and defendants’ motion for sanctions is granted.”

— Illinois Appellate Court, Third District, April 9, 2026

The India Angle

Indian Law Equivalent

Absolute litigation privilege in common-law jurisdictions (including India) means that statements made in judicial proceedings cannot give rise to liability in defamation, civil conspiracy, or similar torts. In India, the principle is embedded in Section 13 of the Defamation Act and in judicial immunity principles derived from the common law. A party who erroneously files a document in court proceedings — even a wrong version of an agreed order — is protected from civil suit in respect of that filing, provided it was made in the course of bona fide judicial proceedings. This privilege is absolute: it applies even where the party or their counsel was mistaken, as long as the filing was not fraudulent in the criminal sense.

Bar Council Rules

BCI Rule 52 (accurate citation of authorities) applies directly to the AI-hallucination misconduct. The attorneys’ fees sanction in this case — awarded for the costs defense counsel incurred in responding to the AI-fabricated brief — has an Indian analogue in Order XXV CPC (security for costs) and the courts’ inherent power to award actual costs as a punitive measure. Indian courts have awarded wasted costs against parties and advocates who file frivolous or misleading submissions that force the opposing party to incur unnecessary legal costs.

Practical Advice for Indian Advocates

• When defending against an AI-hallucinated appellate brief, explicitly move for attorneys’ fees and costs incurred in responding — courts in multiple jurisdictions are now willing to award these as sanction remedies, making AI misconduct financially consequential for the offending party.
• In post-dissolution civil disputes, be cautious about pursuing claims based on conduct that was part of prior judicial proceedings — absolute litigation privilege is a near-absolute bar in common-law jurisdictions including India, and AI research on such claims is prone to hallucinating exceptions that do not exist.
• Illinois Supreme Court AI policy (referenced in the ruling) is an example of a growing number of jurisdictions implementing formal AI governance rules for court filings — Indian advocates should track similar developments by the BCI and the Supreme Court.

Quick Takeaways

• Defense counsel who must respond to AI-hallucinated appellate briefs can seek attorneys’ fees as a sanction — making the AI error directly financially punitive for the offending party.
• An appeal that is both substantively meritless and contains AI-fabricated citations presents courts with the cleanest case for granting comprehensive sanctions.
• Absolute litigation privilege is a fundamental common-law doctrine that AI tools frequently misrepresent or ignore — verify independently whether a contemplated lawsuit is barred by litigation privilege before filing.

Deep Dive: Attorneys’ Fees as the True Cost of AI Misconduct — Beyond Fines and Warnings

The sanctions remedy in Hulvat v. Gumina — attorneys’ fees incurred by defense counsel in responding to the AI-hallucinated brief — represents the most practically significant financial consequence for AI-hallucination misconduct. Most AI-sanction cases to date have involved monetary penalties paid to the court (as in Gentry v. Thompson, where Harris was fined $250 and Roquemore $1,000). These court-payable fines, while meaningful symbolically, do not compensate the opposing party for the actual cost they incurred in dealing with the AI errors.

An attorneys’ fees award, by contrast, is calculated based on the actual time defense counsel spent reviewing the AI-fabricated brief, researching the correct law, drafting responses, and (in some cases) appearing at show cause hearings about the AI errors. In a complex appellate case handled by experienced counsel, these fees can easily exceed $10,000 to $50,000. The Heimkes v. Fairhope case, which involved more egregious misconduct, resulted in a $55,597 attorneys’ fees sanction — reflecting the actual cost to defense counsel of managing the consequences of the AI-hallucinated and otherwise deficient filings over years of litigation.

For self-represented litigants and their AI tools, this financial exposure may not register until a sanctions order arrives. But for law firms and businesses who are tempted to cut costs by relying on AI-generated legal research without adequate verification, the potential attorneys’ fees sanction in the opposing party’s favour should be a powerful deterrent. The cost of one thorough citation-verification step — perhaps two to four hours of a junior associate’s time — is trivially small compared to the cost of a sanctions order requiring payment of defense counsel’s fees for responding to the error.

Indian advocates and their clients should understand that Indian courts’ power to award wasted costs under Order XXV CPC and under the inherent jurisdiction of courts includes the power to award costs that reflect the actual prejudice caused to the opposing party. Where an AI-hallucinated brief forces opposing counsel to spend substantial time identifying and responding to fabricated citations, an application for wasted costs representing those fees would be well-founded and likely to succeed before courts that have become alert to AI-related misconduct.