When Briefs Blur Truth and Technology
A federal appeals court delivered a sharp rebuke that echoed across the legal community. The three judge panel of the 5th U.S. Circuit Court of Appeals ordered attorney Heather Hersh to pay $2,500 after it found she relied on artificial intelligence without proper verification. The sanction arose after the court identified fabricated case citations and serious misstatements within a filed brief.
The court made clear that this episode did not stand alone within recent judicial experience. Judges expressed frustration that AI generated false citations continue to appear in formal filings despite repeated public warnings. The panel stated that the problem shows no sign of abating within federal courts. Such language signaled a deeper concern about professional standards and courtroom integrity.
At the center of the dispute stood a brief that contained invented quotations and distorted legal authorities. The panel discovered twenty one instances that reflected either fabricated language or serious misrepresentation of governing law. This pattern forced the judges to question not only accuracy but candor toward the tribunal. The sanction against Hersh thus represented more than a monetary penalty for isolated oversight. It marked a warning that trust in the judicial system cannot withstand careless reliance on unverified digital output.
A Sanction That Signals Judicial Resolve
The controversy reached the 5th U.S. Circuit Court of Appeals in the case of Fletcher v. Experian Info Solutions. The appeal arose from a lawsuit that accused a lender and a credit reporting agency of violations under the Fair Credit Reporting Act. A federal district judge in Texas had imposed sanctions after he found insufficient pre filing investigation of the client’s claims.
That earlier order required Shawn Jaffer and his firm, then known as Jaffer and Associates, to pay a combined $23,000 in attorney fees to the defendants. The district court concluded that the complaint lacked minimal factual and legal grounding at the time of filing. However, the appellate panel later reversed that sanctions award after its own review of the record. The reversal did not end the matter because concerns about the appellate brief soon surfaced.
Before the reversal issued, the panel identified twenty one fabricated quotations or serious misstatements within the submitted brief. The court responded with a show cause order that required Heather Hersh to explain the discrepancies. That order placed the spotlight on authorship, research methods, and the duty of verification before filing. The judges sought clarity about whether artificial intelligence played a role in the flawed citations.
Jennifer Walker Elrod authored the opinion that addressed Hersh’s response to the show cause directive. She described the explanation as not credible and misleading in several material respects. The opinion stated that Hersh admitted use of artificial intelligence only after a direct question from the court. Elrod indicated that prompt acceptance of responsibility could have resulted in a lesser penalty.
The panel found that Hersh attributed the inaccuracies to public case versions and well known legal databases. Judges rejected that account after they compared cited passages with authoritative sources. The opinion stated that her statements evaded the central issue of independent verification. It emphasized that officers of the court owe candor and accuracy without qualification. The sanction therefore reflected a judicial determination that misleading responses compound underlying citation errors.
Courts Confront a Surge of AI Hallucinations
The Hersh matter fits within a broader national pattern that concerns federal and state courts alike. Judges across jurisdictions report briefs that contain fictitious cases or distorted quotations. What once appeared as a novelty now reflects a persistent challenge to judicial administration.
A database maintained by French lawyer and data scientist Damien Charlotin tracks confirmed incidents of artificial intelligence hallucinations in United States filings. As of this week, the database listed 239 documented cases submitted by attorneys. That tally underscores how quickly reliance on generative tools has outpaced caution.
Appellate judges view these incidents as threats to both ethics and procedure. Courts depend on accurate citations to resolve disputes and maintain consistent precedent. Fabricated authority forces judges and clerks to expend scarce time on verification. Such burdens erode efficiency and strain confidence in counsel representations. The integrity of adversarial advocacy suffers when courts must police basic factual accuracy.
The 5th Circuit confronted these concerns when it considered whether to craft a special rule for generative artificial intelligence use. In 2024, the court evaluated a proposal that would have regulated such tools at the appellate level. Ultimately, the judges declined to adopt a separate rule after internal deliberation. They concluded that existing professional conduct standards already impose adequate duties of competence and candor.
That choice placed responsibility squarely on attorneys rather than on new procedural mandates. The court signaled that ignorance of technological risks no longer qualifies as a plausible excuse. Public reports since 2023 have documented repeated episodes of artificial intelligence citation errors. Judicial opinions now reflect impatience with explanations that shift blame to software or databases. Within this landscape, appellate courts demand vigilance as a basic professional obligation.
The Legal Profession at a Crossroads
These developments place the legal profession at a decisive moment of responsibility. Lawyers must confront how technological tools reshape research habits and courtroom preparation. Courts now signal that competence requires mastery of both doctrine and digital risk.
Verification remains a non negotiable duty of counsel in every filing. No software platform can absolve an attorney from personal review of cited authority. Professional judgment demands careful comparison between generated text and authoritative sources. Legal education must therefore emphasize critical evaluation alongside technical literacy.
Artificial intelligence tools can assist research through rapid synthesis of complex material. Yet such tools cannot replace disciplined analysis or ethical accountability before a tribunal. Credibility in court rests on trust that each citation reflects authentic and verified authority. As technological change accelerates, advocacy will depend on lawyers who combine innovation with unwavering fidelity to truth.
