Federal Rule of Civil Procedure 11(b) requires an attorney who signs a filing to certify, after "an inquiry reasonable under the circumstances," that the legal contentions are warranted by existing law and the factual contentions have evidentiary support. Courts applying Rule 11 to AI-generated filings have been consistent on one point: using an AI tool to draft does not change or lower that certification duty.
How courts have applied Rule 11 to AI hallucinations
In Mata v. Avianca, Inc., No. 22-cv-1461 (S.D.N.Y. 2023), the court found that submitting a brief with fabricated citations, regardless of how those citations were generated, was itself a violation of the reasonable-inquiry standard. The court's reasoning was straightforward: an attorney who signs a filing is representing to the court that a reasonable inquiry was made, and running a citation through an AI tool without independently checking it against a primary source does not satisfy that standard.
What "reasonable inquiry" means in practice
Courts have not required a specific tool or process, but the documented cases converge on a practical minimum: existence must be confirmed against a primary source, quotations must be checked against the actual opinion text, and the case must actually stand for the cited proposition. Falling short of any one of these three has been sufficient to support sanctions in multiple documented cases.
Does the certification duty shift to the AI vendor?
No. Courts have been clear that the signing attorney bears the Rule 11 duty regardless of which tool assisted in drafting, and regardless of whether the tool's own marketing claimed high accuracy. This is consistent with the broader principle that an attorney cannot delegate professional judgment, including verification, to a third-party tool without independent review.
The 2026 case roundup
Beyond Mata, courts have applied Rule 11 or state-law equivalents in Park v. Kim (2d Cir. 2024), Wadsworth v. Walmart Inc. (D. Wyo. 2025), and Coomer v. Lindell (D. Colo.), among many others documented in the Charlotin database. See our full 2026 case roundup for a broader survey.
Expert perspective
Ethics counsel who advise law firms on AI adoption have generally recommended treating AI-assisted drafting the same way firms historically treated junior associate or contract-attorney research: useful, but never filed without a supervising attorney's independent verification of citations before signature.
A practical documentation checklist
- Keep a dated record showing every citation was checked for existence, quotation accuracy, and proposition support before filing.
- Note which tool, if any, was used for drafting and whether it was independently verified.
- If an error is later found, document the correction and any notice given to the court promptly.
A common question
Can Rule 11 sanctions reach a supervising partner who didn't personally run the AI query?
Yes, in principle. Rule 11(b) applies to the person who signs the filing, and courts have applied the certification duty broadly across a filing chain in documented cases. Supervisory attorneys who sign a filing without confirming the underlying research was verified carry real exposure.
Related reading
- Rule 11 Sanctions for AI Hallucinations: The 2026 Case Roundup
- What Happens If You Cite a Hallucinated Case in a Filing
- A Law Firm AI Safety Policy Template You Can Actually Use
- AI-Generated Brief Risk Mitigation: A Framework for Firms
- Mata v. Avianca: What the Sanctions Order Actually Says
Check a brief before you file it →
How Rule 11 sanctions motions actually get litigated
A Rule 11 motion typically follows a specific procedural path: the moving party must generally serve the motion on opposing counsel first and allow a 21-day safe-harbor period during which the allegedly offending filing can be withdrawn or corrected without formal sanctions, before the motion can actually be filed with the court. This safe-harbor mechanism gives attorneys who discover their own AI-hallucination error, or have it pointed out informally, a real opportunity to correct course before facing a formal sanctions motion, which is one more reason prompt self-correction is treated so much more favorably than defending a challenged citation.
Sanctions beyond Rule 11 itself
Courts are not limited to Rule 11 when addressing AI-hallucination incidents. Many have relied on their inherent authority to sanction conduct that abuses the judicial process, which does not require the same procedural safe-harbor as a Rule 11 motion and can be invoked sua sponte by the court itself. State courts often apply their own analogous rules or statutes, which vary somewhat in procedure but generally converge on the same substantive standard: a reasonable inquiry into the accuracy of factual and legal contentions before filing.
What "reasonable under the circumstances" has meant in AI cases specifically
Courts evaluating AI-hallucination incidents under the reasonable-inquiry standard have generally held that awareness of AI hallucination risk is now widespread enough that failing to independently verify AI-generated citations does not meet the standard, regardless of how novel or unfamiliar the specific AI tool used might have been to the individual attorney.
Final takeaway
Understand the safe-harbor mechanism and use it: if you discover an AI-hallucination error in your own filing, correcting it promptly, ideally before any formal sanctions motion is even served, remains the single most effective way to minimize consequences under the Rule 11 framework as courts have consistently applied it.
A common question
Can opposing counsel recover their attorney's fees for responding to a hallucinated citation, even outside a formal Rule 11 motion?
Yes, in several documented cases, courts have awarded fee-shifting specifically tied to the additional time and expense opposing counsel incurred confirming a citation did not exist, separate from any broader sanctions. This is one more reason the cost of skipping verification is rarely limited to the sanctioned attorney's own side of a case.
Keep this safe-harbor timeline in mind specifically: correcting an error within days of it being flagged is treated very differently from correcting it only after a formal motion is served.
The safe-harbor period exists precisely to reward this kind of prompt diligence, and attorneys who use it well tend to face materially lighter consequences than those who wait it out.