The Tenth Circuit's AI-Filing Proposal Turns Human Review Into a Procedural Control
A September 18, 2026 proposed rule from the U.S. Tenth Circuit would require lawyers and self-represented litigants to certify human review of filings prepared with generative AI.
The U.S. Tenth Circuit's September 18, 2026 proposal treats human verification as a procedural control for AI-assisted court filings. Reuters reports that the proposed rule would require lawyers and self-represented litigants to certify that generative-AI-assisted filings were reviewed by a human, with comments open through October 18. Reuters
This is not an AI ban
The proposal does not prohibit drafting with AI.
It changes the responsibility boundary.
The person submitting the filing must certify that the document has been reviewed and that legal citations and factual claims meet the ordinary obligations of court practice.
That is a governance pattern many software teams are already applying to automated code changes.
Verification becomes part of the workflow
The control can be expressed as:
AI assistance → human verification → certification → submission
The legal system therefore turns an abstract "humans remain responsible" principle into an explicit process checkpoint.
Why citations are the obvious pressure point
Generative models can produce plausible but incorrect citations.
Legal documents amplify the cost of that failure because a wrong citation can change the meaning of an argument and damage trust in the filing.
The proposed rule therefore focuses on verification rather than merely disclosure.
Consequences make the control real
Reuters reports that noncompliance could lead to sanctions or other procedural consequences.
That matters because a policy without an enforcement mechanism is only guidance.
What the proposal does not settle
The rule is proposed, not final.
It also applies only to the Tenth Circuit's jurisdiction and procedural environment.
The broader question is whether other courts adopt similar controls, different controls, or no special rule at all.
Observatory interpretation
The strongest signal is not the court's opinion about AI.
It is the conversion of human oversight into an auditable workflow step.
That pattern can generalize to other high-stakes systems where people must remain responsible for machine-assisted decisions.
Sources
Evidence
Sources & further reading
Primary sources, official disclosures, and external research used to ground this report.
- Reuters — US appeals court rule would require lawyers to certify AI-prepared filingsreuters.com
Independent September 18, 2026 reporting on the Tenth Circuit proposal, comment period, and proposed effective date.
- U.S. Court of Appeals for the Tenth Circuitca10.uscourts.gov
Primary court source for the proposed rule, public notices, and final text.
Keep Exploring
Related observations.
Anthropic's Embedded Evaluators Change What Independent AI Oversight Could Mean
The important change is not the investment alone: independent evaluators are being given employee-like access inside a frontier AI company, which creates both stronger visibility and new independence questions.
OpenAI's New Misalignment Framework Turns Rogue Model Behavior Into a Trackable Incident Class
OpenAI is moving model misalignment reporting from occasional research disclosures toward an ongoing incident process. The important signal is the reporting mechanism itself—and its limits.
npm's Stage-Only Tokens Put a Human Gate Between CI and Publication
npm automation can now prepare a release without being able to publish it directly. That small permission change matters because it separates build automation from the final act of putting a package on the public registry.