--- title: "United States federal courts: AI rules for lawyers" url: "https://ailawframework.com/jurisdictions/united-states-federal-courts" type: "jurisdiction" site: "AI Law Framework" id: "US-FED" kind: "federal" country: "US" parent: "United States" last_checked: "2026-10-31" as_of: "2026-10-31" last_reviewed: "2026-11-03" --- # United States federal courts: AI rules for lawyers > United States federal courts: court and regulator AI instruments and the tasks that change here. Canonical page: https://ailawframework.com/jurisdictions/united-states-federal-courts Site index for agents: https://ailawframework.com/llms.txt ## What this means for lawyers There is no nationwide federal court rule on lawyers' use of generative AI. The Judicial Conference's evidence committee has not recommended action on proposed Federal Rule of Evidence 707 (machine-generated evidence); it has revised the proposal and will study it further with deepfakes. Disclosure depends on the court. For example, the Northern District of Texas requires a brief prepared using generative AI to say so on its first page, and filing without that disclosure certifies no generative AI was used. Check local rules and the presiding judge's standing orders in every matter. The Fifth Circuit decided not to adopt a certification rule, reminded counsel of Fed. R. App. P. 46(b)(1)(B) and warned that "I used AI" is no excuse, and relies on existing rules and its inherent power; in February 2026 it sanctioned counsel $2,500 for an unverified AI-drafted reply brief and for lack of candour in responding to the show-cause order. Verify every citation and quotation an AI tool produces before filing. As at 31 October 2026. ## Telling the court and the client **Tell the court:** Varies by court Court disclosure: depends. No general federal rule; some courts require it (for example N.D. Tex. LR 7.2(f)), and the Fifth Circuit and EOIR have none. Before the USPTO there is no general duty to disclose AI use, but disclosure is required where the use is material to patentability (37 CFR 1.56(b)). Client disclosure/consent: no rule found at federal level. EOIR's memo notes that attorneys may need to consult clients about generative AI use, citing ABA Formal Opinion 512. - [Special Order No. 2-98 (2024 Local Rules Amendments)](https://www.txnd.uscourts.gov/sites/default/files/orders/2-98.pdf) - [Court Action on Proposed AI Rule](https://www.ca5.uscourts.gov/news/proposed-ai-rule) - [PM 25-40, Use of Generative Artificial Intelligence in EOIR Proceedings](https://www.justice.gov/eoir/media/1410621/dl?inline) - [Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office, 89 FR 25609](https://www.govinfo.gov/content/pkg/FR-2024-04-11/pdf/2024-07629.pdf) **Tell the client:** Varies Same basis and sources as above. ## Last checked Last checked 31 October 2026. ## What changed in United States federal courts in the last 12 months - 18 February 2026 · Dated: [Fletcher v. Experian Information Solutions, Inc., No. 25-20086 (sanction for unverified AI-drafted reply brief)](#ai-rule-jr-us-fed-ca5-fletcher-v-experian-sanction) ## AI instruments ### Courts and tribunals #### Judiciary interim guidance on AI and AI Task Force (reported after the Judicial Conference session, September 2026) - Type: Guidance for judges · Judicial guideline - Issuer: Administrative Office of the US Courts / Judicial Conference of the United States - Date on page: 17 September 2026 - Effective date: Not stated - Status: 2026-09-17; start date not confirmed - Last checked 31 October 2026 - URL: https://www.uscourts.gov/data-news/judiciary-news/2026/09/17/judiciary-cites-progress-case-management-property-authority-and-ai The federal judiciary's news report on the Judicial Conference session says the AI task force appointed in 2025 by the Director of the Administrative Office has identified more than 60 issues, formed seven subgroups, and that the Director has issued interim guidance to the courts, developed by the task force, with more to follow. Courts were cautioned not to delegate core judicial functions such as decision-making or adjudication to AI, and all judiciary users were reminded they are accountable for work done with AI assistance. This concerns AI use inside the judiciary, not a rule for lawyers appearing before it. Key points: - Courts should not delegate decision-making or case adjudication to AI. - Judiciary users remain accountable for AI-assisted work. - More guidance is to be rolled out as issues are resolved. #### Report of the Advisory Committee on Evidence Rules (proposed new Federal Rule of Evidence 707, machine-generated evidence) - Type: Report, research or submission · Report - Issuer: Judicial Conference of the United States, Advisory Committee on Evidence Rules - Date on page: 17 May 2026 - Effective date: Not stated - Status: 2026-05-17; start date not confirmed - Last checked 31 October 2026 - URL: https://www.uscourts.gov/sites/default/files/document/advisory_committee_on_evidence_rules_may_2026.pdf The Advisory Committee met on 7 May 2026 and, in a report dated 17 May 2026, said it does not recommend action on proposed new Rule 707 at this time. The proposed rule (approved for publication by the Standing Committee in June 2025) deals with machine-generated evidence. The Committee has revised the proposal and plans further study of it, together with the separate problem of deepfakes, at its next meeting. No federal evidence rule specific to AI-generated evidence is in force. Key points: - Rule 707 is not law and the report does not send it to the Judicial Conference for final approval. - The report says the proposed Rule has been revised and will be studied further alongside deepfakes. - The revised draft (set out in the report) is retitled 'Evidence Produced by Artificial Intelligence and Presented at Trial Without an Expert', focuses on 'artificial intelligence' rather than 'machine-generated' evidence, presumes an expert foundation is needed, and adds a notice provision. - In the same report the Committee recommended final approval of Rule 609 amendments, which are unrelated to AI. #### Fletcher v. Experian Information Solutions, Inc., No. 25-20086 (sanction for unverified AI-drafted reply brief) - Type: Judgment - Issuer: United States Court of Appeals for the Fifth Circuit - Date on page: 18 February 2026 - Effective date: 18 February 2026 - Status: 18 February 2026; start date not confirmed - Last checked 31 October 2026 - URL: https://www.ca5.uscourts.gov/opinions/pub/25/25-20086-CV0.pdf Following a show-cause order, the court found that appellant's counsel used AI to draft a substantial portion, if not all, of a reply brief and did not verify the accuracy of the content, and that counsel was not forthcoming in responding to the order. It imposed a $2,500 sanction payable to the clerk within 30 days. The opinion says the court relies on existing rules and its inherent power rather than an AI-specific rule. Key points: - Sanction was for unverified AI-generated quotations, citations and assertions in a brief, and for lack of candour in the response. - The court noted Fed. R. App. P. 46(c) and its inherent power to sanction abuse of the judicial process. - The opinion is a published decision of a federal court of appeals. #### Local Rule LR 7.2(f) and LCrR 47.2(e): disclosure of use of generative artificial intelligence in briefs (Special Order No. 2-98) - Type: Court rule or order · Court rule - Issuer: United States District Court for the Northern District of Texas - Date on page: 10 June 2024 - Effective date: 3 September 2024 - Status: 3 September 2024; start date not confirmed - Last checked 31 October 2026 - URL: https://www.txnd.uscourts.gov/sites/default/files/orders/2-98.pdf A brief prepared using generative AI must say so on its first page under the heading 'Use of Generative Artificial Intelligence'. If the presiding judge directs, the filing party must identify the specific parts prepared using generative AI. A party who files a brief without the disclosure certifies that no part of it was prepared using generative AI. The same text applies to criminal briefs under LCrR 47.2(e). This text is taken from the special order as adopted; check the court's current local rules for later changes. Key points: - Disclosure goes on the first page of the brief under a prescribed heading. - Omitting the disclosure operates as a certification that no generative AI was used. - The definition covers any tool that generates new content such as text or images in response to a prompt, whatever it is called. - Applies to civil and criminal proceedings in the district. #### Court Action on Proposed AI Rule: decision not to adopt a generative AI rule (proposed Fifth Circuit Rule 32.3 amendment) - Type: Court policy - Issuer: United States Court of Appeals for the Fifth Circuit - Date on page: Not stated - Effective date: Not stated - Last checked 31 October 2026 - URL: https://www.ca5.uscourts.gov/news/proposed-ai-rule The Fifth Circuit's announcement says the court has determined not to adopt a rule governing the use of generative AI, and links a notice stating the court 'has decided not to adopt a special rule regarding the use of artificial intelligence in drafting briefs at this time'. The notice reminds parties and counsel of their duties under Fed. R. App. P. 46(b)(1)(B), says filings must be carefully checked for truthfulness and accuracy, and warns that "'I used AI' will not be an excuse for an otherwise sanctionable offense". The court's later Fletcher opinion explains that the rejected proposal (Notice of Proposed Amendment to 5th Cir. R. 32.3, 4 January 2024) would have required counsel and pro se litigants to certify either that no generative AI was used or that a human checked AI-generated text for accuracy, and that the court concluded existing rules were sufficient. Counsel in that court have no AI-specific certification requirement but remain subject to existing rules and sanctions. Key points: - No AI-specific certification applies in the Fifth Circuit as at the date checked. - The court's notice reminds counsel of Fed. R. App. P. 46(b)(1)(B); the Fletcher opinion adds Fed. R. App. P. 46(c) and the court's inherent power to sanction. - District courts within the circuit may have their own requirements (for example N.D. Tex. LR 7.2(f)). ### Government and other bodies #### Policy Memorandum 25-40: Use of Generative Artificial Intelligence in EOIR Proceedings - Type: Government policy or guidance · Government guidance - Issuer: Executive Office for Immigration Review (US Department of Justice), Office of the Director - Date on page: 8 August 2025 - Effective date: 8 August 2025 - Status: 8 August 2025; start date not confirmed - Last checked 31 October 2026 - URL: https://www.justice.gov/eoir/media/1410621/dl?inline General guidance to EOIR adjudicators on parties' use of generative AI in immigration court and Board of Immigration Appeals matters. It says EOIR has neither a blanket prohibition on generative AI nor a mandatory disclosure requirement, but individual adjudicators or courts may adopt standing orders. It says practitioners who file hallucinated or erroneous AI content likely violate professional conduct obligations and may face discipline under 8 C.F.R. 1003.102 (including 1003.102(c), knowingly or with reckless disregard offering false evidence), and that adjudicators who suspect improper use should report it to EOIR's Attorney Discipline Program. The memo creates no enforceable rights, and it says the Department of Justice was expected to issue department-wide guidance by the end of 2025 (not checked). Key points: - No EOIR-wide AI disclosure rule, but local standing orders are permitted. - Practitioners should confirm the accuracy of any citations or research done with generative AI. - Cites ABA Formal Opinion 512 on when client consultation or disclosure may be required. - Refers suspected improper use to the Attorney Discipline Program. #### Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office (89 FR 25609) - Type: Government policy or guidance · Government guidance - Issuer: United States Patent and Trademark Office - Date on page: 11 April 2024 - Effective date: 11 April 2024 - Status: 11 April 2024; start date not confirmed - Last checked 31 October 2026 - URL: https://www.govinfo.gov/content/pkg/FR-2024-04-11/pdf/2024-07629.pdf Federal Register notice (Docket No. PTO-P-2024-0013) reminding parties and practitioners how existing USPTO rules apply when AI tools are used. There is no prohibition on using AI tools to draft documents for the USPTO and no general obligation to disclose their use, but a duty to disclose arises where the use is material to patentability under 37 CFR 1.56(b). Signing a paper certifies a reasonable inquiry under 37 CFR 11.18; simply relying on the accuracy of an AI tool is not a reasonable inquiry, so the signer must review the contents and ensure the statements are true. Practitioners must protect client confidentiality (37 CFR 11.106) when using AI systems and comply with foreign filing licence and export rules. The guidance is not substantive rulemaking and does not have the force of law. Key points: - Applies to practice before the USPTO (patents, trademarks, PTAB, TTAB), not to courts generally. - No general duty to disclose AI use to the USPTO, except where the use is material under 37 CFR 1.56(b). - Reviewing AI output is required: relying on an AI tool is not a reasonable inquiry under 37 CFR 11.18. - Sharing confidential client information with third-party AI systems engages 37 CFR 11.106 and possibly export and foreign filing licence rules. - Follows the USPTO's 6 February 2024 guidance for its judicial boards. #### Federal judge standing orders on generative AI (trackers) - Type: Report, research or submission · Tracker - Issuer: Various US federal judges - Date on page: Not stated - Effective date: Not stated - Last checked 10 October 2026 - URL: https://www.ropesgray.com/en/sites/artificial-intelligence-court-order-tracker - Secondary source - Recorded under United States; shown here because the issuer is in United States federal courts. Law360 and Ropes & Gray track individual standing orders requiring certification or disclosure of AI use (e.g. a W.D.N.C. order). Key points: - Ropes & Gray tracker: https://www.law360.com/pulse/ai-tracker ## Sources - Administrative Office of the US Courts / Judicial Conference of the United States: [Judiciary interim guidance on AI and AI Task Force (reported after the Judicial Conference session, September 2026)](https://www.uscourts.gov/data-news/judiciary-news/2026/09/17/judiciary-cites-progress-case-management-property-authority-and-ai) · last checked 31 October 2026 - Judicial Conference of the United States, Advisory Committee on Evidence Rules: [Report of the Advisory Committee on Evidence Rules (proposed new Federal Rule of Evidence 707, machine-generated evidence)](https://www.uscourts.gov/sites/default/files/document/advisory_committee_on_evidence_rules_may_2026.pdf) · last checked 31 October 2026 - United States Court of Appeals for the Fifth Circuit: [Fletcher v. Experian Information Solutions, Inc., No. 25-20086 (sanction for unverified AI-drafted reply brief)](https://www.ca5.uscourts.gov/opinions/pub/25/25-20086-CV0.pdf) · last checked 31 October 2026 - United States District Court for the Northern District of Texas: [Local Rule LR 7.2(f) and LCrR 47.2(e): disclosure of use of generative artificial intelligence in briefs (Special Order No. 2-98)](https://www.txnd.uscourts.gov/sites/default/files/orders/2-98.pdf) · last checked 31 October 2026 - United States Court of Appeals for the Fifth Circuit: [Court Action on Proposed AI Rule: decision not to adopt a generative AI rule (proposed Fifth Circuit Rule 32.3 amendment)](https://www.ca5.uscourts.gov/news/proposed-ai-rule) · last checked 31 October 2026 - Executive Office for Immigration Review (US Department of Justice), Office of the Director: [Policy Memorandum 25-40: Use of Generative Artificial Intelligence in EOIR Proceedings](https://www.justice.gov/eoir/media/1410621/dl?inline) · last checked 31 October 2026 - United States Patent and Trademark Office: [Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office (89 FR 25609)](https://www.govinfo.gov/content/pkg/FR-2024-04-11/pdf/2024-07629.pdf) · last checked 31 October 2026 - Various US federal judges: [Federal judge standing orders on generative AI (trackers)](https://www.ropesgray.com/en/sites/artificial-intelligence-court-order-tracker) · last checked 10 October 2026 ## Key facts - Kind: Federal - Task versions written for this jurisdiction: 0 - AI instruments recorded: 8 *Practice guidance for planning and review. Confirm current primary sources and the requirements of the actual matter. This is general information, not legal advice.*