This summary is based on the United States Court of Appeals for the Ninth Circuit’s published decision in Doe v. GitHub, Inc., filed September 16, 2026.
The Ninth Circuit panel affirmed the district court’s dismissal of claims brought under the Digital Millennium Copyright Act (DMCA) by a group of programmers against GitHub, Inc., Microsoft Corporation, OpenAI and several affiliated entities. The plaintiffs alleged that GitHub’s AI-powered code writing tools, GitHub Copilot and Codex, reproduce copyrighted code authored by the plaintiffs without proper attribution, violating 17 U.S.C. § 1202(b) of the DMCA.
GitHub Copilot and Codex are artificial intelligence tools that assist programmers by generating code snippets. These tools are trained on millions of public software projects hosted on GitHub, including code published under open-source licenses that often require attribution. The plaintiffs contended that Copilot and Codex sometimes produce code copied from their work but fail to include accompanying copyright management information (CMI), such as author names and license terms, thereby violating the DMCA.
The court explained that § 1202(b) prohibits the intentional removal or alteration of CMI from copies of protected works and the distribution of such works when CMI has been removed or altered without authorization. CMI is defined as information conveyed with copies of a work, including copyright notices and license terms.
The panel first held that the plaintiffs had Article III standing to bring their claims, noting that they adequately alleged a substantial risk of injury from the AI tools reproducing their code without attribution. The court found the complaint’s factual allegations plausible, citing examples where the AI output closely matched plaintiffs’ code, academic research indicating such systems may memorize training data, and GitHub’s implementation of a duplicate-detection feature as evidence that verbatim copies may be generated.
However, the court declined to consider the plaintiffs’ “input” theory of liability, which contended removal of CMI occurred during the AI training process, determining that this theory was forfeited because plaintiffs failed to raise it properly before the district court.
On the merits of the “output” theory, the court agreed with the district court that the plaintiffs failed to state a claim under the DMCA. The court emphasized that Copilot and Codex do not remove or alter CMI from existing copies of protected works but instead generate new works based on statistical patterns learned during training. Because the AI does not produce copies from which CMI is removed or changed, the DMCA’s § 1202(b) protections do not apply.
The court explained its interpretation of the relevant DMCA provisions:
“One who creates a new work and fails to include CMI cannot be said to have ‘removed’ or ‘altered’ anything.”
It further clarified that § 1202(b) protects CMI only when it has been removed or altered from copies of works as fixed in material objects. The court rejected the notion that the plaintiffs’ code was “identical” to the AI’s generated output in the sense needed to infer removal of CMI, finding instead that the AI’s output was better characterized as newly created works that lacked the plaintiffs’ CMI but were not copies from which it had been removed.
The court also rejected the argument that the DMCA should impose liability where an AI tool’s output is substantially similar to existing code but omits CMI, noting that such claims fall within the traditional copyright infringement framework rather than the DMCA’s separate protection for copyright management information.
Accordingly, the Ninth Circuit affirmed dismissal of the plaintiffs’ DMCA claims. Contract claims alleged by plaintiffs remain pending in the district court and were not addressed by the appellate court.
Editorial: Potential Impact
This decision by the Ninth Circuit may have significant implications for the intersection of copyright law and artificial intelligence technologies, particularly those that generate creative content based on large datasets of copyrighted material. By affirming that the DMCA’s prohibitions on removal or alteration of copyright management information do not extend to AI-generated works that do not reproduce existing copies, the court delineates the DMCA’s scope in the context of generative AI.
The ruling suggests that developers and providers of AI tools that create new content modeled on copyrighted materials may avoid DMCA liability for failure to include CMI, so long as the output is not a direct copy of protected works with CMI removed or altered. This may provide some legal clarity and reduce risk for AI developers who rely on large-scale datasets for training.
At the same time, the court’s decision leaves open the possibility that traditional copyright infringement claims, based on substantial similarity, could be pursued for AI-generated content that closely replicates protected works without attribution. The delineation between DMCA claims and copyright infringement claims in such contexts may become a key battleground as AI technologies evolve.
Because this is a published Ninth Circuit opinion addressing novel issues arising from AI and copyright, it could influence future litigation involving AI-generated content, both within the Ninth Circuit and potentially beyond. The decision carefully avoids expanding the DMCA to cover AI outputs in ways that the court views as inconsistent with statutory text, signaling judicial caution in applying existing copyright statutes to emerging digital technologies.
Source: Read the full decision on CourtListener
Court: Court of Appeals for the Ninth Circuit | Docket: 24-7700 | Date filed: 2026-09-16
This article was generated automatically from a public court record and is provided for informational purposes only. It is not legal advice. Refer to the linked source document for the authoritative text of the decision.