Courts are issuing standing orders about AI faster than anyone can track them. Some require disclosure. Some have banned AI-generated filings outright. Most say nothing at all. Here's the actual legal landscape.
When you hear "AI in the courtroom," most people picture a robot on the witness stand or an attorney whispering to an earpiece. The reality is more mundane — and more consequential. The real AI question isn't about gimmicks. It's about whether you can use AI tools to prepare your filings, and what you're required to disclose if you do.
Courts are adapting in real time. As of 2026, there is no uniform federal rule on AI. Individual judges, districts, and circuits are issuing their own standing orders. The rules vary dramatically — and if you're pro se, you're responsible for knowing the rules in your specific court.
Courts have sorted themselves into roughly three groups on AI.
Some courts now require attorneys — and in some districts, pro se litigants — to certify whether any AI tool was used to draft a filing. The Judicial Conference's Committee on Rules of Practice and Procedure began circulating proposed national disclosure rules in 2024. Several circuits moved ahead of that guidance with their own standing orders.
Disclosure orders typically require something like: "Counsel certifies that no generative AI was used to draft this filing without human review and verification of all citations." Variations include requiring disclosure of which tool was used, or certification that every AI-generated statement was independently verified.
What this means for you: If your district has an AI disclosure requirement, you must follow it. This isn't optional even for pro se litigants. Before you file anything, check your local rules and the individual judge's standing orders on the court's website.
A smaller number of judges have gone further, prohibiting AI-generated filings entirely or requiring that any AI-assisted filing be reviewed line-by-line and certified as accurate by the human submitting it. These orders were largely prompted by incidents where attorneys filed briefs containing fabricated citations — cases and quotations that don't exist — generated by ChatGPT without verification.
The most prominent example is Mata v. Avianca (S.D.N.Y. 2023), where attorneys submitted a brief containing six completely fabricated case citations. The judge imposed sanctions. That case became the catalyst for courts nationwide to issue standing orders.
Bans typically don't prohibit AI research tools — they prohibit submitting AI-generated text as your own without verification. The distinction matters: using AI to understand your case is different from having AI write your motion and filing it verbatim.
The majority of courts still have no explicit AI policy. The absence of a rule doesn't mean anything goes. The underlying professional conduct rules still apply: you must verify every citation, every factual claim, every legal proposition before filing. If you use AI and it invents a citation, you're responsible — not the AI.
There are two distinct questions: what AI can do to help you prepare, and what AI can do in the courtroom. They are very different.
This is where it gets complicated. Courts are physical spaces with their own rules, and judges have broad authority to control courtroom conduct.
Phones and laptops: Many courtrooms prohibit electronic devices outright, or require permission to bring them in. If devices are allowed, using them during proceedings typically requires permission.
AI earpieces: Real-time AI earpiece assistance is one of the most contested areas. Some courts have explicitly banned them. Others haven't addressed them. The underlying concern is that an earpiece receiving instructions from an AI model is effectively having a non-attorney represent you — which raises self-representation issues and potential disruption concerns. See our dedicated guide on AI earpiece legality for a deeper analysis.
Referencing AI-generated notes: Bringing printed notes to court is standard practice. Those notes can include AI-assisted analysis. The question of real-time AI consultation in court is different from having done AI-assisted research the night before.
The single biggest concrete risk from AI in legal proceedings is fabricated citations. Large language models can generate plausible-sounding case names, docket numbers, and quotations that do not exist. This happens because the model is predicting what a citation would look like, not retrieving an actual case from a database.
If you file a brief with a fake case citation:
The rule: Every case citation you use must be independently verified. Look it up on a real legal database. Read the actual opinion. Confirm the quotation exists where the AI says it does. This is non-negotiable.
Courts that require disclosure often don't define what counts as "AI-assisted." Does using a spell-checker count? A grammar tool? The concern is clearly about generative AI drafting substantive legal content, not basic word processing. But if you're uncertain, disclose.
The more important principle: the point of AI-assistance disclosure is accuracy. The court wants to know if there's a heightened risk that your filing contains unverified content. If you've verified everything carefully, a disclosure note just clarifies your process — it doesn't create a problem.
The cases that have resulted in sanctions involved attorneys who submitted AI output without reviewing it. That's the failure mode courts are targeting. Careful, verified use of AI as a research and drafting aid is a very different thing.
Most AI disclosure orders were drafted with attorneys in mind. But pro se litigants are still subject to court rules and sanctions. Faretta v. California gives you the right to represent yourself; it doesn't exempt you from the rules of the court. You're held to the same procedural standards as a licensed attorney.
Courts generally give pro se litigants some extra latitude in interpretation when they clearly don't know a rule. But sanctions for filing fabricated citations are unlikely to be waived just because you're representing yourself — the harm to the court is the same either way.
The practical guidance for pro se defendants: use AI to understand your case, prepare your arguments, and draft filings — then verify every factual and legal claim before you submit anything. That process protects you under any AI disclosure framework, whether your court has one or not.
Be My Own Attorney is built for exactly this: AI-assisted case preparation that you control and verify, designed for pro se defendants who need to show up ready.
Join the WaitlistFree while in beta. No credit card required.