A citation can exist in every database and still be completely wrong for your argument.
Most people assume that if a citation is "real" - if the case exists, if the statute is current - then it's correct. That's not how legal citation works. A citation isn't just about existence. It's about whether the authority actually supports the legal proposition it's cited for.
AI tools don't understand this distinction. They'll cite a landlord-tenant statute in a personal injury case because the statute mentions "damages" and the brief is about damages. The connection is superficial, but the format is perfect.
Subject-matter mismatches - Florida Statute § 83.49 (residential security deposits) cited in a commercial lease dispute. Both involve landlords. Both involve money. But they govern entirely different legal relationships, and the residential statute doesn't apply to commercial tenancies.
Entity-type errors - Rule 1.140(b) subsections for corporate defendants cited in a case involving an individual. The rule text references "the defendant" generically, but specific subsections apply only to certain entity types.
Wrong court hierarchy - A First DCA opinion cited as binding authority in a case before the Fourth DCA. The case is real, the holding is relevant, but it's only persuasive - not binding - in that jurisdiction.
Superseded authority - A case that was good law when decided but has since been overruled, distinguished, or superseded by statute. AI models trained on older data won't know the current status.
Why this matters: A judge who catches an inapplicable citation won't just ignore it - they'll question every other citation in your filing. One bad citation undermines the credibility of your entire argument.
Westlaw and LexisNexis can tell you whether a case exists and whether it's still good law. But they won't tell you whether the case supports the specific argument you're making. That requires reading the opinion and understanding the context of your brief.
This is exactly the gap AI Detector Pro fills. Beyond verifying existence, ADP runs applicability checks that flag:
• Subject-matter mismatches between the cited authority and your document's legal context
• Entity-type conflicts (individual vs. corporate, residential vs. commercial)
• Court hierarchy issues for Florida jurisdictions
• Stale authority warnings for older cases
ADP's Deep Analysis tier takes applicability checking to the next level. Multiple independent AI models (Claude and GPT) read the actual opinion text and your document's context, then independently assess whether each citation supports the argument it's cited for. Results are aggregated via consensus voting - if the models disagree, the citation is automatically flagged for human review.
When a citation doesn't support your argument, Deep Analysis searches for alternative authorities that would. Instead of just telling you something is wrong, it helps you fix it.
