10th Circuit proposes AI filing certification rule
The U.S. Court of Appeals for the 10th Circuit proposed a rule requiring lawyers and self-represented litigants to certify human review of AI-assisted legal filings, establishing one of the most concrete governance frameworks yet for agent use in federal courts.
The Certification Requirement
Under the proposal, litigants would be required to certify that any document prepared with AI assistance has been reviewed by a human "who has independently verified the existence of all cited authorities; the accuracy of all quotations, citations, and legal analysis; and the document's compliance with all relevant rules and ethical obligations," according to Reuters reporting on September 18, 2026.
The rule targets a specific vulnerability in agent-assisted legal work: the ability of large language models to fabricate case citations, misquote precedent, and misrepresent legal analysis—failures that have already triggered attorney sanctions in other jurisdictions. By requiring independent human verification before filing, the 10th Circuit is attempting to create an accountability layer between agent output and court submission.
Comment Period and Timeline
Reuters reported that the Denver-based court is accepting public comment on the proposal until October 18, 2026, and plans to finalize the rule by December 1, 2026. This compressed timeline—roughly 2.5 months from proposal to adoption—suggests the court views the governance gap as urgent, particularly as AI tools for legal research and drafting proliferate among both large firms and solo practitioners.
Why This Matters for the Agent Economy
The 10th Circuit's proposal represents a concrete regulatory constraint on how agents can be deployed in a major institution. Unlike hypothetical "AI regulation," this rule creates documented, enforceable requirements that will shape how legal agents are built, marketed, and used across a multi-state federal jurisdiction.
The certification requirement also establishes a model that other courts and regulators may adopt: rather than banning agent use outright, create mandatory checkpoints that preserve human accountability. This approach allows agents to remain productive tools while reducing the risk of agent-generated fraud or malpractice.
The rule does not specify penalties, fines, or enforcement mechanisms beyond existing attorney discipline processes, leaving open questions about how violations will be detected and sanctioned. Whether the certification requirement will be self-policed or subject to court auditing remains unclear pending finalization.