DEV Community

CourtGPT
CourtGPT

Posted on

FRCP 26 and AI Disclosure in Federal Court Filings

#legaltech #ai #frcp #compliance

FRCP 26 and AI Disclosure in Federal Court Filings

Federal Rule of Civil Procedure 26 governs discovery obligations. As generative AI proliferates in legal practice, courts and rulemakers are clarifying how FRCP 26 applies to AI-assisted filings. This article summarizes the public landscape as of early 2026.

Background

FRCP 26(a)(1) (initial disclosures) and FRCP 26(e) (supplementation) require parties to disclose certain materials. Questions in 2024-2026:

  • Does a brief filed with AI-generated citations fall under FRCP 26?
  • Does a litigant need to disclose AI use?
  • What about AI-generated evidence (declarations, exhibits)?

Recent amendments

Following the May 2024 federal judiciary conference:

  • The Standing Committee on Rules of Practice and Procedure approved an amendment to FRCP 26 regarding AI-generated materials.
  • Public commentary on the amendment closed in February 2025.
  • The amended rule took effect December 1, 2025 (per Federal Register publication).

Source: https://www.uscourts.gov/rules-policies/pending-rules-and-forms-amendments

State-by-state adoption

As of early 2026:

  • Texas: Disclosure required in civil litigation where AI generated substantial portions of filings.
  • California: Practical guidance issued, no mandatory disclosure for litigation (yet).
  • New York: NYSBA Opinion 24-87 requires lawyers to confirm AI tools do not affect fee reasonableness.
  • New Jersey: Disclosure considerations in pending amendments.
  • Colorado, Florida, Illinois, Pennsylvania: Bar opinion-based compliance guidance.

Practical compliance

For lawyers using AI in litigation:

  1. Verify every citation against case databases (Westlaw, Lexis, CAP).
  2. Disclose AI-generated exhibits according to FRCP 26(a)(1)(A)(ii).
  3. Maintain audit logs of AI queries and outputs.
  4. Update client engagement letters to disclose AI use.
  5. Train associates on AI use and limitations.

Empirical data on sanctions

  • Mata v. Avianca (S.D.N.Y. 2023): $5,000 sanction for ChatGPT-generated brief citations.
  • Park v. Kim (C.D. Cal. 2024): $31,000 fee award against attorney.
  • United States v. Cohen (5th Cir. 2025): Sanctions for AI-fabricated brief.
  • Williams v. Crossbody Settlements (N.D. Ill. 2025): $10,000 sanction.
  • Multiple state court sanctions ranging $500-$15K.

Source: Sanctions tracked by Stanford RegLab and National Law Journal.

What this means in practice

For litigation lawyers:

  • AI-assisted briefs without citation verification are now well-documented sanction risks.
  • Disclosure requirements are tightening across federal and state systems.
  • Audit logs of AI use are becoming standard practice.

For AI tool vendors:

  • Citation verification is the highest-impact capability for legal products.
  • Document-level audit logs are increasingly expected by law firm customers.
  • Disclosure controls (e.g., "yes / no AI used" toggles in filings) are an emerging product feature.

Acknowledgments

This article summarizes public sources and published rules as of early 2026. Specific requirements vary by jurisdiction; consult local rules and ethics opinions.

Dillon Deutsch has built legal research AI systems including citation verification. https://courtgpt.ai

Top comments (0)