Appellate AI Litigation Explosion Hinges on Three SCOTUS Opinions

Five years ago, writing a legal treatise describing “an appellate approach to AI litigation strategy” felt innovative. Today, the US appellate courts—and appellate practitioners—are front and center in resolving the most important questions of our technological age.

Why is the appellate approach driving artificial intelligence litigation, and what does it mean for winning these critical cases?

Along with the large increase in the frequency and scope of AI litigation, much of the current litigation turns on readings of three US Supreme Court decisions. The legal novelty and high stakes of these disputes is bringing the appellate courts to the table.

Let’s begin with some data. Since 2016, 69 federal appellate opinions have mentioned AI. Forty-one of those AI opinions were issued in 2024 and 2025. When human judgement is applied to determine whether an appellate ruling involves AI in a legally relevant sense, the number of rulings is 27. The US Court of Appeals for the Ninth Circuit has seven opinions, and the Federal Circuit has five. The Second Circuit has four, as does the Fourth. The DC Circuit has three AI opinions. The First, Third, Sixth, and Seventh Circuits each have one AI opinion. The Fifth, Eighth, Tenth, and Eleventh Circuits don’t yet have an AI opinion.

Why so much appellate AI litigation? First, and most straightforward, is the increasing frequency of litigation involving AI. In 2020, there were 21 district court decisions. In 2024, there were 184.

There is also the wide sweep of AI legal areas. AI is affecting every part of our lives that law governs. At the appellate level, so far, the opinions had involved seven different legal areas: the Communications Decency Act, securities, trade secrets, privacy, products liability, Section 1983, and the First Amendment.

Second, the stakes of AI litigation are often enormous. The results will determine the delicate arrangement of innovation and creativity, of national security and economic interests.

There is also a more subtle driver of the appellate role in AI litigation: the centrality of US Supreme Court cases to resolve novel legal questions. We can see this most clearly in the ongoing AI copyright litigation. There are 45 pending cases involving copyright and AI.

The district courts currently hearing these cases are spread across the First, Second, Third, and Ninth Circuits. And in each case, most of the briefing is about the application of the same set of Supreme Court legal decisions.

Regardless of where the litigation is occurring, the focus is on Oracle, Warhol, and Feist. Many of the central pleadings in these cases could have been filed in any of the other cases with only modest refinements.

To an unusual extent, this copyright district court litigation is in substance appellate litigation.

Consider the recent hearing in the Kadrey v. Meta Platforms matter before Judge Vince Chhabria in San Francisco. The first attorney to speak at the hearing was Kannon Shanmugam, an appellate attorney who has argued 150 appeals.

To guide the argument, Chhabria began with the following: “[I]f you wouldn’t mind beginning the discussion by just focusing on the issue in general without regard to any of the facts in this case.”

Chhabria’s focus on only the law, and not the facts (at a summary judgment hearing no less), reflects his background as a former law clerk to Justice Stephen Breyer, and chief of the Appellate Section in the San Francsico Attorney’s Office. The balance of the argument involved application of appellate cases, hypotheticals, and waiver.

Another example is the pending Thomson Reuters v. ROSS Intelligence matter, in which the authors of this article are representing the AI innovator. Judge Stephanos Bibas, who sits on the Third Circuit, presided over the trial litigation by designation. Like Chhabria, he also managed the summary judgment litigation as if it were an appeal—in fact, he reversed himself.

In his first opinion, Bibas concluded there was a factual dispute over the copyrightability of the training data and whether the AI innovator’s use of the data was fair. And in that opinion, he signaled that Oracle was the most applicable precedent. But days before trial, Bibas changed his mind and asked the parties to rebrief the copyright issues.

A few months later he issued an opinion coming to the opposite conclusions on copyrightability and fair use—this time Warhol, not Oracle, governed. Ultimately, Bibas acknowledged the issues were hard and certified the case to be heard on appeal under 28 USC Section 1292(b). The case will likely be argued in the spring of 2026.

A third high-profile AI litigation matter also has significant appellate dimensions. In a widely covered ruling, Judge William Alsup held that while Anthropic PBC’s use of copyrighted material was fair, its use of pirated material may not be.

In a surprising twist, the parties recently agreed to a proposed settlement. And one—if not the—principal reason the plaintiffs’ gave for reaching a settlement is the risk of appellate proceedings.

AI litigation depends critically on district court litigation. The cases involve complex and extended depositions, detailed expert reports, and lengthy hearings. But as the Meta hearing, Ross proceeding, and Anthropic filings illustrate, the appellate perspective is joining the trial teams at critical moments.

To mimic the AI marketing slogan of a leading global technology company, the moment of Appellate Intelligence has arrived.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Mark Davies is a partner at White & Case and the author of the fourth edition of “Artificial Intelligence: Law & Litigation” (2025).

Kufere Laing is an associate at White & Case, representing clients in appellate litigation and dispositive trial court motions in federal and state courts across the country.

Davies and Laing, along with Yar Chaikovsky, Anna Naydonov, and Andy LeGolvan, are litigating an AI copyright appeal discussed in this article.

Laetitia Haddad contributed to this article.

Write for Us: Author Guidelines