The Byron White U.S. Courthouse in Denver, home of the U.S. Court of Appeals for the Tenth Circuit. Photo: Nyttend/Wikimedia Commons (public domain).
After the Fifth Circuit tried and retreated in 2023, the Tenth Circuit is now advancing a proposed rule for mandatory human verification of AI-assisted court filings. Comments on the rule are due by October 18th. If passed, it would take effect in January 2027. The rule would make both lawyers and pro se litigants certify that a human independently checked any AI-assisted filing before submission. The certification would become part of the normal procedure for giving documents to the court. The proposed rule would allow for striking filings that don't comply and even waiving affected issues as part of its sanctions provisions.
As courts grapple with how to treat AI-assisted legal work, the technology has changed dramatically since the Fifth Circuit considered and rejected a similar rule in 2023. AI is now far more capable and far more present in everyday legal work, including from pro se litigants. The Tenth Circuit’s proposal would put it at the forefront of federal appellate rulemaking on AI-assisted filings, even as lower courts and state courts have already adopted myriad rules of their own.
Open Models
Z.ai’s Beijing headquarters. Photo: Z.ai.
The Chinese company behind one of the most popular open-weight AI models — GLM — is under fire for grabbing users’ code. Beijing-based Z.ai makes ZCode, a coding assistant similar to Codex and Claude Code that can read and work across an entire software project. Developers recently discovered that a feature turned on by default was uploading entire projects to Alibaba’s cloud without permission. Z.ai apologized, disabled the feature and said the uploaded data had been deleted, but the number of affected users and projects remains unknown. The fallout forced Z.ai to release ZCode’s source code, enable zero-data retention and promise a full security report.
Quotable
“It’s not the technology alone that determines whether AI helps or harms.”
— Adam Gelb, president and CEO, Council on Criminal Justice
In Other News
A Tennessee woman is suing Fargo for $10 million over a facial recognition lead that wrongly put her in jail for six months. (ABC News)
Legal tech startup Legora is working on making AI work better with legal authority and citations. (Legora)
E-discovery vendor Relativity is rebranding its platform around AI, with RelativityOne becoming Relativity aiR. (Relativity)
Less than a third of law firms surveyed said they have changed pricing for AI-assisted work. (BigHand)
California is speeding up implementation of new AI oversight laws, while New York also advances its own framework. (California Governor’s Office; NYDFS)
Shoddy software during the 2025 California bar exam led to a proposed $1.35 million settlement. (Reuters)
Google’s new laptops run Android and are built around AI. (Google)
The Gates Foundation is backing a project to improve AI access for speakers of languages underserved by today’s systems. (Gates Foundation)
About Plain Thinking
Grant Morgan is an attorney, former prosecutor and former Tribune Publishing reporter who covered Gary, Indiana, and Chicago’s northern suburbs. Plain Thinking covers AI and consequential technology affecting lawyers, police and courts.
Every Tuesday and Friday · Get Plain Thinking by email · LinkedIn
If an issue includes paid placements, they’ll be clearly labeled. Companies don’t pay for editorial inclusion.
