Does New York require attorneys to disclose AI use in court filings? Partially

Does New York require attorneys to disclose AI use in court filings? Partially. 22 NYCRR Part 161, adopted March 2026 and effective June 2026, expands attorney certification duties to cover AI-generated content — but it doesn't require you to stamp 'AI-assisted' on the face of your filing. It's a verification mandate, not a disclosure mandate.

The distinction is critical for managing partners building compliance protocols. New York chose to strengthen the existing certification framework rather than create a new disclosure regime. Attorneys must verify that filings contain no fabricated content — including AI hallucinations — but the mechanism is certification, not labeling.

What 22 NYCRR Part 161 Actually Requires

Part 161 amends the existing attorney certification rules to explicitly address AI-generated content. When signing a filing, attorneys certify that they've verified the accuracy of all content, including any portion generated or assisted by artificial intelligence tools. This extends to case citations, statutory references, factual assertions, and legal arguments. The rule doesn't ban AI use — it makes the attorney personally responsible for catching AI errors before they reach the court.

Who Part 161 Applies To

The rule applies to all attorneys admitted to practice in New York state courts. This includes attorneys of record, supervising attorneys, and any attorney whose signature appears on a court filing. It does not apply to pro se litigants, court staff, or judicial officers — those groups fall under separate governance frameworks. For firms, the practical impact falls hardest on signing partners and associates whose names appear on filings. If a junior associate uses AI to draft a brief and a partner signs it, the partner bears the certification obligation.

Penalties and Enforcement

Sanctions are available under Part 161 for attorneys who certify filings containing fabricated AI content. The court can impose monetary sanctions, strike pleadings, issue adverse inference instructions, or refer attorneys to disciplinary authorities. New York courts have already demonstrated willingness to sanction attorneys for AI-related failures — multiple reported cases involved attorneys submitting AI-hallucinated citations before Part 161 even took effect. The new rule gives courts explicit statutory authority to do what they were already doing under inherent powers.

How New York Differs from Federal Court AI Rules

Federal courts in New York have been more aggressive. The Southern District (SDNY) and Eastern District (EDNY) have individual judge orders requiring affirmative AI disclosure on the face of filings. Part 161 deliberately stops short of this — it requires verification but not disclosure. This means an attorney filing in SDNY might need to disclose AI use, while filing the same motion in New York Supreme Court requires only certification that the content is accurate. Managing partners with practices spanning both systems need to track which standard applies where.

Compliance Playbook for New York Firms

Start with workflow documentation. Every AI-assisted filing needs a verification step before the signing attorney certifies it. Build a citation-checking protocol — run every case citation through Westlaw, Lexis, or another verified legal database. Implement a review layer where a human attorney reads AI-generated arguments for logical coherence and factual accuracy. Train associates on what Part 161 requires and what it doesn't — they can use AI freely, but they own the output. Finally, keep records of your verification process. If a court questions a filing, you want documentation showing the steps you took.

The Bottom Line: New York's Part 161 requires attorneys to verify AI-generated content is accurate when certifying filings, but doesn't mandate disclosing AI use on the face of the document — build verification workflows, not disclosure stamps.