Skip to content
Judicial AI Standards Institute sealJudicial AI Standards Institute

The Bench & Bar AI Brief | Issue 003 | June 9, 2026

The Court's AI Rule Is Still the Same: Verify Before You File

New court orders and rules show the same pattern: judges are not punishing AI use as such, they are punishing unverified filings, false citations, false quotations, and candor failures.

Back to BriefOpen checklist

Tennessee Edition

Welcome to Issue 003 of The Bench & Bar AI Brief, Tennessee Edition.

The strongest AI lesson for judges and lawyers this week is plain: courts are not treating AI as the offense. They are treating unverified legal work as the offense.

That distinction matters. It keeps the conversation out of panic and inside ordinary professional duty. A lawyer may use new tools, but the signer still owns the filing. A court may see AI in the background, but the question in front of the court is familiar: are the authorities real, are the quotations accurate, and was counsel candid when the problem surfaced?

The Ninth Circuit made that point in LNU v. Blanche. The court sanctioned two attorneys after appellate briefs included nonexistent cases, misattributed quotations, and serious misstatements. The order did not turn on the mere use of generative AI. It turned on the filing of inaccurate work and the failure to correct the record with candor.

Florida is moving the same rule into statewide court procedure. Its amended Rule 2.515 requires signers of court filings to represent that the legal authorities identified in a filing exist and are accurately cited. The rule applies to attorneys and unrepresented parties, and sanctions may follow after notice and an opportunity to be heard.

The professional-discipline side tells the same story. In Oklahoma, reciprocal discipline followed an AI-fabricated-citation matter that had already produced public discipline in Texas, Tennessee, and Alabama. That is a practical warning for Tennessee lawyers: an AI citation problem may not stay inside one case or one jurisdiction.

The self-represented-litigant cases are also growing. Georgia warned a pro se appellant about fictitious cases and quotations associated with irresponsible generative AI use. Ohio affirmed refusal to set aside sanctions where a Tennessee resident blamed DeepSeek and limited access to Ohio legal research tools for citation problems.

Bench & Bar takeaway.

The safe standard is not "never use AI." The safe standard is "never file what you have not verified." Courts can teach that rule without adopting a complicated technology code.

A practical courtroom version.

  1. 1Check every case outside the AI tool.
  2. 2Confirm every quotation against the source.
  3. 3Verify the rule statement, not just the citation.
  4. 4Keep a record of the verification path.
  5. 5Correct mistakes quickly and candidly.

Why it matters.

That is not a new duty. AI just makes the old duty easier to break at scale.

Three questions for local discussion.

  1. 1Should Tennessee courts address AI filing problems through statewide rule text, local orders, bench-bar education, or ordinary sanctions practice?
  2. 2Should law offices require a written verification step before AI-assisted research or drafting reaches a filed document?
  3. 3How should courts explain AI risks to self-represented litigants without creating fear or blocking access to helpful tools?

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