The Southern District of Ohio sits in Columbus, Cincinnati, and Dayton -- three cities driving Ohio's financial services, insurance, and technology sectors. This court handles everything from class actions to patent disputes, and with the University of Cincinnati and Ohio State feeding a robust legal talent pipeline, AI adoption among practitioners here is accelerating fast. Judge Michael Newman has already drawn a line.
AI Disclosure Rules in the District of Ohio, Southern
The Southern District of Ohio has no district-wide AI order. The AI rules come from individual judges, and they point in different directions. Judge Michael J. Newman (Western Division, Dayton) bans attorneys and pro se parties from using AI in preparing any filing: the AI provision of his Standing Order Governing Civil Cases, effective August 27, 2025, carves out legal search engines such as Westlaw or Lexis, internet search engines, and Microsoft Office-type products such as Word, and it requires everyone to tell the Court immediately if they discover AI use in any document filed in their case or an opponent's. (Judge Newman's Standing Order Governing Civil Cases, section VI (ohsd.uscourts.gov, read Oct 21, 2026)) Judge James L. Graham (Eastern Division) takes the opposite approach: his order lets parties use generative AI but requires a declaration captioned 'Disclosure of the Use of Generative AI' that identifies the AI-assisted content by page, section, or paragraph, names the platform used, and certifies that the filer reviewed the source material, verified accuracy, and complied with Rule 11(b). (Judge Graham's Standing Order on the Use of Generative AI, October 21, 2025 (ohsd.uscourts.gov, read Oct 21, 2026))
Judge Newman's order is notable because it goes further than disclosure: it prohibits the use, so disclosing after the fact does not cure it. The order says its purpose is not to stop lawyers from using legal search engines or Microsoft products but to prevent counsel and pro se parties from citing false, misleading, or hallucinated cases and law. Judge Graham's order is the conventional disclose-and-certify model.
The <a href="https://www.nycbar.org/reports/artificial-intelligence-in-federal-courts-a-random-sample-survey-of-judges/" target="_blank" rel="noopener">March 2026 NYC Bar study</a> found that 41.7% of the federal judges surveyed (112 respondents) reported no official AI use policy, counting 17.6% who discourage AI use without formally prohibiting it. It surveyed judges, not districts, so it says nothing specific about the Southern District of Ohio. What matters here is that two judges have written orders, so attorneys filing in Columbus, Cincinnati, or Dayton should check their assigned judge's rules before using AI at all.
Individual Judge Standing Orders
Judge Michael Newman's standing order is the most restrictive AI measure in the Southern District. It goes beyond disclosure: using AI in preparing a filing is itself a violation (the carve-outs cover legal search engines, internet search engines, and Microsoft Office-type tools), not just a failure to disclose or verify. Judge James L. Graham's October 21, 2025 order sits at the other end: AI is allowed, but any use to help generate content in a filing must be disclosed in a declaration and verified.
Both orders rest on the logic that courts have applied to outsourced legal work for decades: you can delegate research, but you cannot delegate responsibility. Whether your paralegal wrote the brief, a contract attorney drafted it, or ChatGPT generated it, the attorney of record owns every word.
Individual judges elsewhere have issued AI orders, and we have not verified a national count. In the Southern District, attorneys should check each judge's individual rules page, as additional judges may have adopted requirements since this writing. The court's website and the clerk's office are your best resources for current standing orders.
Key AI Cases in SDOH
Press reports describe sanctions in Ohio federal court for AI-fabricated citations, including one reported in February 2026 involving two attorneys (we have not read the opinions). The cautionary tales from other circuits apply directly as well. Mata v. Avianca in the Southern District of New York remains the landmark -- an attorney submitted fabricated citations generated by ChatGPT and faced sanctions that ended his career as he knew it.
These are not abstract warnings. Ohio attorneys who rely on unverified AI output are playing the same game those sanctioned lawyers played. Judge Newman's order says its purpose is to prevent counsel and pro se parties from citing false, misleading, or hallucinated cases and law.
What Attorneys in SDOH Should Do
Read your assigned judge's AI rules in full. If your case is before Judge Newman, his civil standing order bans AI use in preparing filings, with narrow carve-outs. If it is before Judge Graham, file the 'Disclosure of the Use of Generative AI' declaration whenever AI helped generate any content. Either way, that is a stricter or different standard than many other districts impose.
Build verification into your workflow, not after it. Where AI is allowed, as in Judge Graham's cases, do not draft with AI and then skim the output before filing. Build a step-by-step process: generate, verify every citation on Westlaw or Lexis, confirm factual accuracy, then finalize. The verification is the work.
Consider disclosing even before judges without AI orders. Not every SDOH judge has an explicit AI order yet. Voluntary disclosure shows good faith and protects you if rules tighten mid-case, but it does not cure a violation of a ban: before Judge Newman, the question is whether to use AI at all.
Use enterprise-grade legal AI, not consumer chatbots. Platforms built for legal work include guardrails that consumer tools lack. If you are using free ChatGPT for case research in a federal court, you are one hallucinated citation away from a career-altering sanctions motion.
Keep records of your AI interactions. Save your prompts, the AI output, and your verification notes. If opposing counsel challenges your filing, you want a paper trail showing exactly how you used the tool and how you verified the result.
The Bottom Line
The Southern District of Ohio has no district-wide AI rule, and its two known judge orders differ: Judge Newman bans AI in preparing filings, and Judge Graham requires a disclosure declaration and verification. Know which judge you are before you draft.
Because the rules are judge-specific, check the assigned judge's chambers page for current standing orders at the start of every case. AI compliance in Columbus, Cincinnati, or Dayton is not optional and not aspirational. It is the current reality. Build your process accordingly.
AI-Assisted Research. This piece was researched and written with AI assistance, reviewed and edited by Manu Ayala. For deeper takes and the perspective behind the research, follow me on LinkedIn or email me directly.