The Florida Supreme Court new AI-related filing rule takes effect June 15, and if you read some of the media coverage, you might think that Florida’s courts are entering some kind of futuristic legal era.
This is not the case.
What the court did was remind lawyers of something that predates computers, the Internet, and probably most of the courthouses we practice in today: If you put your name on a record, you are responsible for what’s in it.
That’s the story here.
The rule change comes after courts across the country have been faced with a growing number of filings containing bogus cases, bogus citations and citations that appear perfectly legitimate until someone tries to find them. Many of these problems are caused by lawyers relying too heavily on generative AI tools without verifying the results carefully enough before filing their documents.
And to be honest, some examples are astonishing.
Not because the technology made mistakes. Anyone who has spent time with these systems understands that they can produce errors. What is surprising is how many lawyers apparently trusted the results enough to file them in court without independently verifying them.
This is what judges react to.
Florida’s new rule amends Rule 2.515 to require attorneys signing court documents to certify that the cited legal authorities actually exist and are accurately represented. Courts are also expressly authorized to impose sanctions for violations.
Concretely, this means that lawyers now have even less room to argue: “the software gave me incorrect information”.
But I suspect most judges have felt this way before.
What the Florida Supreme Court really seems to be doing here is standardizing expectations across the state. Different circuits had begun issuing their own administrative orders regarding AI use, disclosures, certifications, and verification requirements. If you have practiced in multiple jurisdictions, the rules may seem inconsistent depending on where your case was pending.
There is now at least one statewide standard.
Honestly, it probably helps everyone.
The biggest issue, however, isn’t really whether lawyers are using AI. Most lawyers already use technology constantly. Legal research platforms themselves have been integrating AI-based tools for years. Young lawyers, in particular, naturally experiment with new systems because they are fast, efficient and often genuinely useful.
The problem begins when efficiency quietly replaces judgment.
This is where things can go wrong.
One thing that non-lawyers may not fully appreciate is how credible AI hallucinations can appear in legal writings. These systems generally don’t spew nonsense that immediately seems ridiculous. They generate quotes that resemble real cases. They produce citations written in compelling legal language. Sometimes the formatting is impeccable.
Until someone checks the source.
That “until someone checks” part matters more than ever now.
Judges should not have to spend time determining whether the cited authority exists. Opposing counsel should not have to waste billable hours chasing down phantom cases. Courts are already moving slowly enough without adding forensic review of citations to motion practice.
And credibility, once damaged before a judge, is difficult to repair.
Any experienced litigator knows this.
You can lose an argument and recover. You can lose a move and recover. But once a court starts to question whether your documents are trustworthy, the problem becomes much bigger than a bad record.
That’s why I don’t consider this rule anti-technology. I also don’t think this means the courts are trying to ban AI from legal practice. This ship sailed a long time ago.
The best comparison is probably GPS.
GPS is extremely useful. Almost everyone depends on it. But if your GPS tells you to drive into a lake, at some point you still have a responsibility to look through the windshield.
AI works the same way.
Used carefully, it can save time and help lawyers organize information more quickly. Used carelessly, it can quickly create a mess.
And the legal profession is probably still in its early stages of determining where those boundaries lie.
A few years ago, many lawyers dismissed concerns about hallucinatory quotes as overblown. At the time, some of these things seemed more theoretical than real. That has changed. Courts across the country now deal with these issues regularly enough that judges clearly feel the need to send a stronger message.
Florida is not alone in this.
Ultimately, what matters most about this rule is not the language itself, but what it says about the direction the courts are taking. Judges become less patient with the idea that AI mistakes are somehow different from other professional failures.
From the court’s perspective, a non-existent cause citation remains a non-existent cause citation, whether the attorney invented it personally or copied it from a chatbot.
And honestly, that’s probably the right approach.
Technology evolves. Professional responsibility is not.
In the end, no software signs the file. No algorithm appears before the judge during a hearing. No chatbot is sanctioned by the court.
The lawyer does it.
This part hasn’t changed at all.
Matthew T. Christ is a partner of Rafferty Domnick Cunningham & Yaffaa preeminent Florida law firm based in Palm Beach Gardena, West Palm Beach, Pensacola and Jacksonville.
