“AI Slop” Is Not a Legal Standard: What Lisandrillo v. Palozzi Asks of Every Lawyer

By Jay Kotzker

Beyond the Hallucination Era

For the past three years, the profession’s conversation about generative AI in litigation has largely been a conversation about fake cases. Lawyers cited opinions that did not exist, courts sanctioned them, and the lesson seemed simple enough: check your citations.

Lisandrillo v. Palozzi, No. 4D2026-2262 (Fla. 4th DCA Sept. 16, 2026), marks a turn in that conversation. Florida’s Fourth District Court of Appeal did not find a single invented case. The citations were real. The court’s concern was something harder to audit and far more fundamental: whether a trained advocate had actually exercised independent judgment in building the arguments placed before it.

That shift, from verifying authority to demonstrating judgment, deserves the attention of every lawyer who drafts, supervises, or signs court filings, whether or not they use AI tools at all.

What the Fourth District Did

The underlying matter is a contested dissolution of marriage case out of Broward County. Petitioner’s counsel sought extraordinary relief in the Fourth District through a petition for writ of prohibition, accompanied by a motion to disqualify the trial judge. Press reports describe the petition as roughly 61 pages and the motion as roughly 33.

In an opinion authored by Judge Robert Gross and joined by Judges Spencer Levine and Caroline Shepherd, the court denied the petition and wrote separately to address what it called “AI slop.” Several features of the opinion stand out:

• The court expressly separated this problem from hallucinations. In a footnote, it stated that the issue had “nothing to do with hallucinated law or false legal citations,” drawing on a January 2026 Bloomberg Law commentary that described filings in which the cases exist but the reasoning does not.
• It identified what it viewed as indicia of non-human authorship. These included unsupported factual assertions about the trial judge that the court could not locate in the record, seemingly random bolded and italicized sentences, and passages the court characterized as legal gibberish. It suggested some arguments may have been generated from mistranscriptions in a rough-draft transcript.
• It named a pattern it called “issue churning.” The court described a laundry list of convoluted arguments that, in its view, no competent lawyer exercising independent judgment would advance.
• It hedged on authorship, then declared authorship irrelevant. The court acknowledged that some portions seemed too unpolished to be machine-generated. Ultimately, it said, whether the filings were AI-generated, AI-assisted, “or something else makes no difference,” because an attorney must stand behind every submission.

The court then ordered counsel to show cause within ten days, “without the use of AI”, why sanctions, including possible referral to The Florida Bar, should not be imposed for frivolous filings. No sanction has been imposed as of this writing.

The Response: “AI Slop” Is Not a Legal Standard

Counsel did not go quietly. Even before her formal response, she publicly stated that she does not use AI to generate court filings, and noted that the court had not accused her of citing fabricated law.

Her 84-page response to the show-cause order, filed this week, reportedly presses that point further. As described in early coverage and commentary, she argues that:

• The legal strategy, arguments, and professional judgment in the filings were her own.
• Every authority was reviewed and verified before filing.
• A self-audit of roughly 237 citation instances surfaced two errors, an outdated rule subsection and a quotation that was not verbatim, which she characterizes as ordinary human mistakes rather than fabricated law.
• The court is, in effect, treating prose that looks AI-generated as evidence of sanctionable conduct, and “AI slop” has no footing as a legal standard.

Whatever one thinks of the underlying filings, that last argument is serious. It reframes the dispute from a question about one lawyer’s work product into a question about how courts may reason from the appearance of a document to conclusions about the process that produced it.

The Real Question: Judgment Is the Standard, Not Authorship

Read carefully, the Fourth District’s opinion and counsel’s response may be less opposed than the headlines suggest. Both, in the end, place the weight on the same thing: the lawyer’s independent professional judgment.

The court said as much when it declared that authorship “makes no difference.” Frivolous arguments, unsupported assertions about the record, and scattershot issue lists have always been sanctionable, long before any chatbot existed. If the filings fall short of that baseline, the outcome would be the same had they been drafted on a typewriter. Counsel, for her part, rests her defense on the claim that the judgment was hers and that she stands behind each argument.

The friction lies in the inferential path. Speculating about AI in a published opinion carries real reputational consequences, particularly when the court itself concedes uncertainty. There is a meaningful difference between two propositions:

1. These arguments are frivolous and unsupported by the record, and counsel is responsible for them.
2. These arguments read like machine output, and that appearance itself suggests counsel failed to exercise judgment.

The first is a familiar, workable standard grounded in existing rules of professional conduct and sanctions law. The second risks turning style into evidence, and as AI-assisted drafting becomes ordinary, “sounds like AI” will describe an enormous share of competent, carefully reviewed legal writing. Bold sentences, confident tone, and long issue lists are not new; neither is bad brief-writing.

The lasting value of Lisandrillo may be the reminder that courts already possess the tools they need. The question is never what wrote this; it is who answers for it, and did they think it through.

What This Means in Practice

Citation verification is now the floor, not the ceiling. For lawyers and legal teams, whether or not AI is part of the workflow, the case points toward a few disciplined habits:

• Curate the issues. Selecting the strongest two or three arguments is itself an act of judgment, and courts read it that way. A long list of loosely connected grounds invites the inference that no one did the choosing.
• Tie every factual assertion to the record. Allegations about a judge’s conduct, in particular, should be anchored to a transcript page or docket entry and checked against the certified transcript, not a rough draft.
• Edit for reasoning, not just accuracy. Ask whether each authority actually advances the argument in which it appears, and whether a skeptical reader can follow the chain from premise to relief.
• Use emphasis sparingly. Formatting should guide the court to what matters most. Indiscriminate bold and italics now read as a warning sign.
• Document your process. A record of who researched, drafted, reviewed, and verified each filing is increasingly valuable both for internal quality control and in the event a court questions how a submission was prepared.
• Adopt a clear AI-use policy. Firms should define approved tools, required human review, disclosure obligations under applicable local rules and orders, and supervisory responsibilities. Florida courts in particular have been active in this area, so check current statewide rules and local administrative orders before filing.

None of these practices is new. Lisandrillo simply makes clear that courts will notice when they are missing and may draw conclusions about why.

Closing Thoughts

At Holon Law Partners, we view AI as a powerful instrument in the hands of a thoughtful lawyer, and never a substitute for one. The tools will keep improving. What cannot be automated is the decision about which argument matters, what the record truly shows, and whether a position is worth a court’s time.

Lisandrillo v. Palozzi is unresolved, and the Fourth District’s ruling on the show-cause response will merit close attention. However it ends, the message for the profession is already clear: a signature on a filing is a promise of judgment. Our clients, and the courts we appear before, deserve nothing less.

This article is provided for general informational purposes only and does not constitute legal advice or create an attorney-client relationship. For guidance on AI governance, litigation practice, or professional responsibility questions specific to your situation, please consult qualified counsel.