Skip to content
Judicial AI Standards Institute sealJudicial AI Standards Institute

The Bench & Bar AI Brief | Issue 007 | July 7, 2026

AI Use Now Needs a Filing Record

A federal standing order, a New York disclosure bill, Florida circuit orders, and bar guidance point to the same rule: AI can help, but the human record must be clear.

Back to BriefOpen checklist

Tennessee Edition

Welcome to Issue 007 of The Bench & Bar AI Brief, Tennessee Edition.

Planned publication date: July 7, 2026.

The first AI warnings for lawyers were easy to understand: do not file fake cases. That warning still matters. But this week's verified source set points to the next stage. When AI touches legal work, courts and lawyers are being pushed toward a record that answers four plain questions: Was AI used? What did it help prepare? Who checked it? Who owns the final filing?

The clearest example is a Middle District of Florida standing order from Judge Moe. The court page and linked order identify a requirement that every filing contain a certification stating whether artificial intelligence was used in preparing the filing. That is not a Tennessee rule, and it should not be described as a district-wide rule unless a later source supports that. But the procedure matters. It turns AI use from an invisible drafting choice into a filing-level certification.

New York is testing the same idea in proposed legislation. Assembly Bill A08546 would add a new CPLR provision and amend appellate briefing rules to require disclosure when generative AI is used in filings. The verified bill text and NYSBA analysis show the core idea: a separate affidavit disclosing AI use and certifying that a human reviewed the source material and verified the generated content as accurate. That bill is proposed, not enacted law. Still, it shows where the pressure is moving.

Florida gives practical court-level examples. The Florida Bar Journal's July and August article, Hallucinated Law: Candor and Competence in the Age of Artificial Intelligence, ties AI use back to ordinary lawyer duties of candor and competence. The verified article says lawyers may use AI only while complying with existing ethical duties and must verify the thoroughness and accuracy of AI-produced research or drafting. Separate administrative orders from Florida's 17th and 19th Judicial Circuits show how local courts are putting that idea into disclosure, certification, independent verification, and sanctions language. Those are Florida circuit examples, not Tennessee law. The lesson travels because the duties are familiar.

The enforcement record explains why a filing record matters. In Landberg v. City of New York, the New York Appellate Division sanctioned an attorney and law firm after a brief prepared with generative AI assistance included nonexistent cases, fictitious quotations, and misrepresentations about real cases. In Delaware's Leiske v. Kidd, the Court of Chancery issued a show-cause order after defendants identified fictitious citations, fabricated quotations, and hallucinated legal propositions in an answering brief. The Delaware matter is not a final sanctions order. Its value here is narrower: the court treated candor and reasonable inquiry as nondelegable duties.

The record should answer four questions.

  1. 1Was AI used?
  2. 2What did it help prepare?
  3. 3Who checked it?
  4. 4Who owns the final filing?

Source signals.

  1. 1Judge Moe's Middle District of Florida standing order turns AI use into a filing-level certification.
  2. 2New York Assembly Bill A08546 would require disclosure and human-verification affidavits for AI-assisted filings if enacted.
  3. 3Florida Bar Journal guidance ties AI use to existing lawyer duties of candor, competence, and verification.
  4. 4Florida circuit orders show court-level disclosure, certification, independent verification, and sanctions language.
  5. 5Landberg and Leiske show why invisible AI work creates visible court-record risk.

A starter filing-record checklist.

For Tennessee courts, lawyers, agencies, and vendors, the question is not whether every jurisdiction has the same rule. They do not. The practical question is whether the file could prove responsible use tomorrow morning.

  1. 1Record whether AI was used in drafting, research, editing, translation, summarizing, discovery, or analysis.
  2. 2Identify the tool, user, date, purpose, and general input category without exposing confidential or sealed material.
  3. 3Keep the prompt, output, or review note when the AI-assisted work may later be challenged.
  4. 4Verify all legal authorities, quotations, procedural claims, citations, and factual assertions against primary or reliable sources.
  5. 5Separate proposed law, advisory guidance, local orders, standing orders, and binding statewide rules.
  6. 6Require a human reviewer to sign off before a filing, order, client advice, public notice, or agency explanation leaves the office.
  7. 7Train staff and vendors that AI may assist the work, but it never holds the legal duty.

Bench & Bar takeaway.

This is not a rule against useful tools. It is a rule against invisible tools. AI can help with legal work only if the human record stays clear. The bench-and-bar standard should be simple: if the work becomes public, official, adversarial, or relied on by a court, the responsible human should be able to show the verification trail.

The question for this week: if an AI-assisted paragraph were challenged six months from now, could your file prove who checked it and what sources they used?

Distribution

Keep up between issues.

Subscribe to the weekly Brief by email, and follow Judicial AI Standard on X for daily AI and courts updates when verified information is available.

Subscribe to the BriefFollow on X
Return to The Bench & Bar AI Brief archive