Thursday, September 17, 2026
Page 5
Ninth Circuit:
Claims Accusing OpenAI of Copying Code Rightly Dismissed
Opinion Says Programmers Failed to State Claim Under Digital Copyright Act Based on Allegations That Generative AI Models Pushed Out ‘Nearly Identical’ Code
By a MetNews Staff Writer
The Ninth U.S. Circuit Court of Appeals held yesterday that a group of programmers, who accused the California-based technology giants OpenAI and GitHub of creating artificial-intelligence generated “output” that is “essentially verbatim” copies of their protected code without providing the required attributions, failed to state a claim under a law designed to capture digital-age copyright violations that may otherwise slip through the cracks of federal regulation.
Yesterday’s opinion is the first appellate-level decision addressing how a provision of the Digital Millenium Copyright Act (“DMCA”), found at 17 U.S.C. §1202(b), applies to generative-AI outputs.
That section provides that “[n]o person shall, without the authority of the copyright owner” “intentionally remove or alter any copyright management information” (“CMI”) or “distribute, import for distribution, or publicly perform works, copies of works, or phonorecords, knowing” that CMI “has been removed or altered.” CMI is defined as including the “information identifying the work” and the “name of…the author.”
In the second-amended pleading at issue in yesterday’s opinion, the plaintiffs alleged that they “published Licensed Materials on GitHub,” the world’s largest hosting platform for open-source software, or that for which creators choose to make the code publicly available on the condition that the author’s name and a copyright notice be included in any derivative use. They further asserted:
“In June 2021, GitHub and OpenAI launched Copilot, an AI-based product that promises to assist software coders by providing or filling in blocks of code using AI. Copilot ignores, violates, and removes the Licenses offered by thousands—possibly millions—of software developers, thereby accomplishing software piracy on an unprecedented scale. Copilot outputs text derived from Plaintiffs’ and the Class’s Licensed Materials without adhering to the applicable License Terms and applicable laws.”
Claim Under DMCA
Circuit Judge Eric Miller authored yesterday’s decision, declaring:
“[W]e agree with the district court that the complaint does not state a claim under the DMCA. The complaint describes an AI tool that does not look up and reproduce stored work but rather creates new work. That new work may or may not infringe plaintiffs’ copyrights in their code, but it cannot reasonably be described as a copy of that code from which CMI has been ‘removed’ or ‘altered.’ ”
The programmers, who identified themselves only as “Doe” plaintiffs, filed the putative class action against the two companies, which are both headquartered in San Francisco, and related entities in November 2022. In the second-amended pleading, they asserted claims for violations of §1202(b) as well as breach of contract causes of action based on the terms of licenses that purportedly accompany the open-source disclosures.
They originally sought “statutory damages” for each “direct violation” of §1202, in an amount they alleged would exceed $9 billion, but dropped the allegation in the subsequent pleadings. In the second-amended complaint, they prayed for injunctive relief, attorney fees, and damages for harms resulting from the purported licensing breaches.
Motions to Dismiss
On June 24, 2024, District Court Judge Jon S. Tigar of the Northern District of California granted the defendants’ motions to dismiss as to the DMCA claim, with prejudice, under Federal Rule of Civil Procedure 12(b)(6), finding that the plaintiffs had failed to cure pleading deficiencies that were highlighted in an earlier order. He wrote:
“The [second-amended complaint] includes the same allegations that Defendants’ programs released, or ‘output,’ code published to GitHub by Does 1, 2, and 5….In its last order, the Court concluded that these facts were ‘not sufficient for a Section 1202(b) claim’ because they were not identical….Because these facts have not changed, the Court must again conclude that Plaintiffs have failed to meet the DMCA’s identicality requirement.”
Tigar declined to dismiss the breach of contract claims, which remain pending.
Miller noted:
“Plaintiffs have articulated two different theories of DMCA violations. The first, which we will call the ‘input’ theory, is that defendants violated section 1202(b)(1) at the training stage by removing CMI from putative class members’ code before feeding the stripped code into Copilot as training data. The second, which we will call the ‘output’ theory, is that Copilot sometimes returns memorized training data as output to users without including the source code’s CMI, and that in doing so Copilot ‘remove[s]’ or ‘alter[s]’ CMI on copies of plaintiffs’ protected work….”
Output Theory
Saying that “[w]e decline to consider the input theory because plaintiffs forfeited it” by not correcting Tigar when he made it clear that he believed they were only pursuing an “output” theory, he turned to the latter theory and compared the plaintiffs’ allegations to the text of the statutory scheme. He opined:
“CMI is ‘information conveyed in connection with copies…of a work,’ not in connection with excerpts or derivative works….To state a claim under section 1202, plaintiffs must…allege that defendants removed or altered CMI from copies of existing protected works; merely alleging that a similar or derivative work does not include the CMI, without facts showing its removal or alteration, is insufficient.”
He said that Tigar’s characterization of this principle as the “identicality” requirement “is something of a misnomer” because the law “does not require literal identicality between the plaintiff’s work and the allegedly infringing work.” The jurist explained:
“To the extent identicality is relevant, it is because one way to determine whether CMI has been ‘remove[d]’ from a copyrighted work is to compare the allegedly infringing copy with the original. If the two works are otherwise identical…but the allegedly infringing version omits the CMI that accompanied the original, a factfinder may reasonably infer that the defendant removed CMI from the copied work.”
Not Identical
Miller agreed with the plaintiffs that “two works need not be literally identical to support an inference that CMI has been removed” but said that the fact “does not help” them “because they have not alleged removal of CMI from a copy of their work, even one with minor alterations.” He pointed out:
“According to the complaint, Copilot relies on a ‘complex probabilistic process’ to predict ‘the most likely solution to a given prompt,’ based on ‘the solution it has found in the most projects’ answering similar questions. Sometimes, plaintiffs allege, that output may match snippets of ‘code from the training data’ [derived from open-source software].”
The judge continued:
“But their account of the algorithm’s internal process…does not describe an action taken with respect to CMI attached to an existing work. Instead, it describes a process through which Copilot generates new works….If Copilot functioned like a search engine and produced outputs that were identical to plaintiffs’ code but did not contain CMI, then plaintiffs might have a stronger claim that Copilot had removed their CMI.”
Recognizing that “Copilot’s output may in some cases be substantially similar to existing code,” he reasoned:
“We express no view on whether that similarity would allow plaintiffs to assert a claim for copyright infringement. But we note that many copyright cases involve the creation of a work that is substantially similar to the plaintiff’s without attribution….If that were all it took to violate section 1202(b), the DMCA would supplant traditional copyright protections and subject defendants to potentially ruinous liability under the DMCA’s enhanced statutory damages…. We decline plaintiffs’ invitation to transform run-of-the-mill copyright infringement claims into DMCA claims.”
The case is Doe v. GitHub Inc., 24-7700.
In a May 2023 order, Tigar ruled that the plaintiffs could proceed pseudonymously after their attorneys, with the Bay Area firm of Joseph Saveria Law Firm LLP, provided threatening emails they received in connection with the litigation, including one that said:
“[G]o…cry about [G]it[H]ub…[I] hope your throat gets cut open and every single family member of you [sic] is burnt to death.”
Copyright 2026, Metropolitan News Company