AI in Attorney Fee Petitions: Tenth Circuit Rules AI Use Alone Won’t Reduce a Fee Award

For the past few years, firms and parties have folded generative AI into research and drafting while a quieter question went unanswered: When a lawyer uses AI and then bills for the time, can the other side cut the fee on that basis alone? On October 30, 2026, the Tenth Circuit Court of Appeals appears to be the first federal court of appeals to squarely address the issue. In Flux Drive Inc. v. Rong, it answered “no” … at least not without proof. And in getting there, the court did something more useful than announce a special rule for AI: It declined to make one.

How The AI Fee Question Arose

The AI fees issue surfaced in the context of an ordinary fee fight. The plaintiffs brought a Colorado trade-secrets suit (CUTSA and the federal DTSA) and named, among others, a set of Illinois defendants (the “Rong Defendants”) as alleged “go-betweens.” Those defendants won dismissal for lack of personal jurisdiction, and that dismissal triggered Colorado’s mandatory fee-shifting statute, Colo. Rev. Stat. § 13-17-201, which entitles a defendant to “reasonable attorney fees” when a tort action is dismissed out on a Rule 12(b) motion. The district court awarded the full lodestar requested: $114,663.11. Supplemental fees for the fee litigation pushed the total to $183,653.61, and the Tenth Circuit noted that appellate fees may yet be added under the same statute. The plaintiffs appealed, and among their challenges to the amount was the argument that matters here: that defense counsel’s use of AI should have shrunk the bill.

The Holding: AI Use Alone is Not a Discount

The plaintiffs argued that defense counsel “should have billed fewer hours due to their reliance on artificial intelligence (‘AI’) and other online research tools.” The Tenth Circuit panel was unpersuaded and affirmed the award in full. The rule the court applied is the one it already uses for every fee dispute, now pointed at a new fact: AI use, standing alone, does not render billed hours unreasonable. A party that wants an AI-related reduction must come forward with evidence that the hours actually billed were excessive. Pointing at the presence of the technology does not carry that burden.

The Court’s Reasoning

The court’s analysis has three steps worth separating, because together they explain why this is a win for recoverability rather than a one-off.

First, the court conceded the plaintiffs’ premise in theory. It acknowledged that a firm’s use of AI and online research tools “may ‘[]reduce the number of hours required’ for research and thereby lower the lodestar amount,” citing the Second Circuit’s Arbor Hill for the unremarkable point that research efficiencies can cut compensable time. Efficiency, in other words, is a legitimate lodestar consideration.

Second, it put the burden where it belongs. The court found the plaintiffs had no proof, stating that they “identify no evidence that counsel’s use of AI resulted in excessive billing.” A generalized assertion that the hours were too high is, as the court put it quoting Anchondo, “an insufficient basis upon which to disturb the district court’s lodestar fee determination.”

Third, it supplied the counterweight that blunts the efficiency argument: Verification is not optional. “[E]thical rules require counsel to verify the cases cited and to ensure that all arguments are grounded in legal authority.” The panel cited three recent AI-hallucination and Rule 11 decisions: Amarsingh v. Frontier Airlines (10th Cir. 2026), Wadsworth v. Walmart (D. Wyo. 2025), and Coomer v. Lindell (D. Colo. 2025) to make the point that whatever time AI saves on the front end, counsel must still spend time confirming the output. That verification labor is part of what legitimizes the hours. So the bottom line: Even assuming AI makes lawyers more efficient, “Plaintiffs have not shown that defense counsel’s use of AI rendered the hours that they actually expended unreasonable.”

The record ultimately carried the award across the line. Defense counsel tracked time to the tenth of an hour, did not block-bill, reduced hours where appropriate, and absorbed extra work when the plaintiffs amended their pleading. On that record, the lodestar presumption held.

What This Means for Practitioners

The practical significance is that the Tenth Circuit folded AI into ordinary lodestar doctrine instead of carving out a special regime. The questions are the familiar ones: reasonable hours, reasonable rate, billing judgment, and AI is just another fact the parties litigate inside that frame.

For fee applicants who use AI, the lesson is clean and complete record-keeping. The award survived because the time entries could withstand a lodestar review: granular increments, no block billing, voluntary reductions. Counsel who use AI should expect opposing counsel (and their own clients) to ask what the tools did and did not do and should keep records that answer it.

For the party challenging a bill, the lesson is that indignation is not evidence. “They used AI, so it should have taken less time” will not move a court. The challenge has to identify specific entries, quantify the claimed excess, and tie it to the hours actually billed which is the same showing required to attack any fee request. Absent that, the lodestar presumption does the work.

Key Take Aways of the Flux Drive Decision

The Flux Drive ruling rests heavily on burden and the plaintiffs’ failure of proof, so its weight on the deeper question, i.e., should AI efficiency reduce a fee? wasn’t reached. The court expressly left that door open: It allowed that AI “might thereby reduce the hours necessary to litigate a case” and faulted the plaintiffs only for not building the record to show it had. A challenger who builds a case against the fees might bring in expert declarations, task-level analysis, and a concrete comparison of AI-assisted time to what the task reasonably required may get a reduction in a future case.

The larger takeaway is almost reassuring in its modesty. Faced with a genuinely new tool, the Tenth Circuit reached for established principles and found they still worked. AI did not break the lodestar; it just became one more thing the parties have to prove.

Mike Laszlo is a Colorado attorney who speaks regularly on attorneys’ fees in Colorado and serves as a retained expert in attorney fee disputes, advising counsel and parties on the reasonableness of fees, lodestar analysis, and fee recovery under Colorado and federal law.

This publication is intended for general informational purposes only and does not constitute legal advice or a solicitation to provide legal services.  The information in this publication is not intended to create, and receipt of it does not constitute, a lawyer-client relationship.  Readers should not act upon this information without seeking professional legal counsel.  The views and opinions expressed herein represent those of the individual author only and are not necessarily the views of Clark Hill PLC.  Although we attempt to ensure that postings on our website are complete, accurate, and up to date, we assume no responsibility for their completeness, accuracy, or timeliness.