On 2 November 2026, the US Court of Appeals for the Tenth Circuit affirmed a fee award of about $114,663 in favour of defendants represented by Brownstein Hyatt Farber Schreck, after Flux Drive's trade secrets lawsuit against a former engineer had been dismissed by a Colorado federal court. Flux Drive, ordered to pay the defendants' legal costs, argued on appeal that the firm's use of AI and online research tools should have reduced the hours, and therefore the bill. Writing for a unanimous panel, Judge Carolyn McHugh rejected the argument: the plaintiffs had not shown that counsel's use of AI made the hours actually spent unreasonable, and they pointed to no specific inflated, duplicated or improper billing entry. The panel accepted that AI can shorten research time, but noted that lawyers must still verify their authorities and arguments. The ruling matters because it is one of the first appellate answers to a question every client is now asking: if the machine did part of the work, why am I paying for it?

In my view, the court got the method right. Fee-shifting review has always been entry-by-entry, and a party that wants a reduction must show where the time was wasted; a general assertion that "AI exists" is not evidence, and courts should not presume that technology was used efficiently, or used at all. I also welcome the emphasis on verification, because the real cost of AI-assisted work often sits in checking what the tool produced. I believe, however, that the decision leaves the harder question open: it does not say that AI efficiency can never justify a lower award, and future challengers with better records, for instance showing that hours were billed for tasks a tool completed in minutes, may succeed. The practical lesson is twofold. Firms should keep time records that show human judgment, not machine output; and clients should negotiate AI use and pricing at the engagement stage, rather than litigating it after the bill arrives!

Source: Reuters, 2 November 2026; Above the Law, 4 November 2026; JD Journal, 3 November 2026. Image: Wikimedia Commons.