Their motion looks convincing. Every citation formatted perfectly. But are the cases real?
Attorneys are using ChatGPT, Claude, and other AI tools to draft motions, responses, and briefs. Some are transparent about it. Most aren't. That's not speculation. It's happening right now, in every practice area, at every level.
The question isn't whether opposing counsel is using AI. It's whether they bothered to verify what the AI gave them.
AI drafting tools are fast. They produce clean formatting, confident language, and citations that look right at first glance. The problem is that "looks right" and "is right" are two very different things. And the attorney who filed that motion may not know the difference.
You can't always tell by looking. That's the whole problem. AI-generated citations are formatted correctly. The volume numbers make sense. The reporter abbreviations are right. The page numbers fall within plausible ranges. Everything looks legitimate.
But there are patterns worth watching for. Cases you can't find on Westlaw, LexisNexis, or any public database. Decisions from the right jurisdiction with holdings that seem too perfectly tailored to the argument being made. Statute sections that exist but govern something completely different from the proposition they're cited to support.
The uncomfortable truth: A fabricated citation in a well-written brief is harder to catch than a real citation in a poorly written one. AI doesn't make formatting mistakes. It makes factual ones.
The more polished the brief looks, the less likely anyone is to question the underlying citations. That's what makes AI-generated filings dangerous. They pass the eye test every time.
Judges take this seriously. In Mata v. Avianca, attorneys were sanctioned after submitting a brief containing ChatGPT-fabricated case citations. The court didn't treat it as a minor oversight. It was a credibility problem that overshadowed the entire filing.
Courts across the country are implementing AI disclosure requirements. Some require attorneys to certify that AI-generated content has been verified. Others require disclosure of AI use in the drafting process. The trend is clear: courts expect citations to be real, and they're not going to be forgiving when they aren't.
If you can demonstrate that opposing counsel filed fabricated citations, that's not just a problem for the citations in question. It's a credibility problem for their entire argument. Judges remember. It can change the outcome of a motion, and it can change how the court views everything that attorney files going forward.
Strategic advantage: Finding a single fabricated citation doesn't just undermine that one point. It gives you grounds to challenge every citation in the filing. If they didn't verify one, why would the court trust they verified any of them?
Upload the PDF of their motion or brief to AI Detector Pro. That's it. Every citation gets extracted automatically and verified against authoritative sources.
Case citations are checked against CourtListener's database of millions of court opinions and Google Scholar's legal database. If the case doesn't exist in either source, you'll know.
Statute citations are checked against the complete Florida Statutes index. All 637 chapters, over 24,800 sections, and 745 part divisions, scraped directly from flsenate.gov. Not just chapter-level checks. Section-level verification that catches invented section numbers, repealed provisions, and wrong-chapter citations.
Rule citations are checked against all 78 Florida Rules of Civil Procedure. Every rule number is verified for existence.
You'll have results within minutes. Green means verified. Red means not found. Amber means it needs a closer look. No manual cross-referencing required.
Existence is just the first question. A citation can be real and still not support the argument it's being used for. That's actually the more common problem with AI-drafted filings. The cases and statutes exist, but they don't say what opposing counsel claims they say.
ADP's Deep Analysis runs each citation through two independent AI models. Both are given the actual source material: opinion text from CourtListener for cases, official catchlines and chapter context for statutes, verified rule numbers for procedural rules. They independently assess whether the cited authority actually supports the legal proposition it's being used for.
If opposing counsel cited a residential tenancy statute in a commercial lease case, Deep Analysis catches it. If they cited a case about personal injury in a breach of contract motion, Deep Analysis catches that too. When the two models agree, you can be confident. When they disagree, the citation gets flagged for closer review.
Think of it this way: The Verification Report tells you whether the citation is real. Deep Analysis tells you whether it actually means what they say it means. Both matter. But the second one is where opposing counsel's AI-drafted arguments tend to fall apart.
If you find fabricated or inapplicable citations, you have options.
File a motion to strike. If citations in opposing counsel's brief are fabricated, move to strike the portions of the brief that rely on them. A brief built on fake authority isn't a brief worth considering.
Raise it at oral argument. Sometimes the most effective moment to surface a fabricated citation is in open court. Ask opposing counsel to direct the court to the cited authority. If they can't, the judge will draw their own conclusions.
Request sanctions. Under Fla. Stat. § 57.105, the court may award reasonable attorney's fees to the prevailing party if the losing party's claims were not supported by the material facts or the existing law. Filing a brief with fabricated citations fits squarely within that framework.
The point isn't to be adversarial for its own sake. It's to make sure the court isn't making decisions based on fake law. If the other side's argument depends on citations that don't exist or don't say what they claim, the court needs to know that. Your client's outcome depends on it.
