TL;DR: The Ninth Circuit on September 16, 2026 affirmed the district court’s dismissal of DMCA Section 1202(b) claims in Doe v. GitHub, Inc., holding that AI-generated outputs do not constitute “remov[al] or alter[ation]” of copyright management information in a way that violates the DMCA. The court left undecided the related input theory, which the plaintiffs forfeited in the proceedings. For trial teams, this decision narrows a key DMCA-based lever in AI-litigation and pushes plaintiffs toward traditional copyright infringement or contract theories, while also shaping how discovery and evidence will be approached in AI-related disputes. This guidance arrives from the U.S. Court of Appeals for the Ninth Circuit on an interlocutory appeal arising from the Copilot and Codex ecosystem used with GitHub and OpenAI technologies.
What happened
In DOE v. GitHub, Inc., the Ninth Circuit examined whether defendants violated 17 U.S.C. § 1202(b) by allegedly removing or altering copyright management information (CMI) in Open Source code used to train or power AI tooling like GitHub Copilot and Codex. The panel held that the DMCA does not reach AI-generated outputs that create new works rather than copying existing works with CMI removed or altered. The court stated that Copilot and Codex “do not ‘remove or alter’ CMI from a copy of an existing protected work but instead create new works that never contained that information” and thus do not satisfy the § 1202(b) “remove or alter” requirement. The Ninth Circuit thus affirmed the district court’s dismissal of the DMCA claim. The opinion further explained that the district court properly declined to consider the input theory because the plaintiffs forfeited that argument by not pressing it in the lower court. The decision was issued on September 16, 2026. See 24-7700 (9th Cir. Sept. 16, 2026) and the published PDF available from the Ninth Circuit’s official site. (cdn.ca9.uscourts.gov)
Key excerpts from the opinion confirm the central rationale: the DMCA’s “remove or alter” language targets CMI that has been removed or altered in existing copies, not the production of entirely new outputs generated by AI models trained on those inputs. The court’s analysis emphasizes that while the outputs may raise copyright concerns, they are not the same as a direct removal or alteration of CMI within a copied work, which is the core of § 1202(b). The ruling thus forecloses a broad “output theory” under § 1202(b), at least in the form presented in this case. The district court’s denial of relief on the DMCA claim was therefore affirmed. (cdn.ca9.uscourts.gov)
Practical implications for trial lawyers
Narrowing the DMCA leverage in AI disputes: This decision signals to litigants and courts that DMCA § 1202(b) claims are unlikely to prevail when the central allegation is that AI outputs reproduce or resemble code without attribution. The Ninth Circuit explicitly rejected the notion that AI-generated outputs automatically implicate CMI under § 1202(b) simply because the outputs do not include licensing or attribution information. For trial teams, this means a strategic reallocation of AI-related claims away from § 1202(b) toward other theories of liability. Expect courts to scrutinize the factual record closely to separate training data issues from the actual outputs when crafting claims. (cdn.ca9.uscourts.gov)
Forensic approach to AI disputes: Because the Ninth Circuit declined to entertain the input theory at this stage, plaintiffs cannot ballast claims on how training data was handled unless those arguments survive in separate procedural tracks. Practitioners should anticipate a continued emphasis on traditional copyright infringement, licensing, trade secret, or contract theories where viable, rather than relying on DMCA 1202(b) as a catchall for AI training and outputs. In defense, emphasize the novelty of AI outputs and the absence of a “copy-without-CMI” scenario within the output itself. (cdn.ca9.uscourts.gov)
Discovery and evidence considerations: The ruling reduces DMCA-based leverage in discovery related to CMI and training data in AI cases, at least for claims framed around output parity with protected code. Attorneys should still consider broader discovery tools for training materials, but the DMCA angle may no longer be a robust gatekeeper for obtaining training data. Litigants should plan discovery around standard infringements, licensing documents, and any contractual obligations governing use of open-source code, rather than seeking broad CMI disclosures under § 1202(b). (cdn.ca9.uscourts.gov)
Strategic shift for case teams: In light of this decision, litigants should assess whether parallel or alternative claims may yield relief, such as copyright infringement claims based on actual copying of protected code, misappropriation theories, or breach of license terms. Consider also the potential role of equitable remedies or injunctions tied to ongoing use of AI tools in a manner that implicates relied-upon licenses. This shift does not eliminate DMCA considerations entirely, but it alters their centrality in many AI-litigation scenarios. (cdn.ca9.uscourts.gov)
Implications for Objection Academy users
Practical training for AI-related evidentiary issues: The Ninth Circuit’s decision underscores the importance of rigorous evidence readiness when AI topics arise in trial settings. Objection Academy’s objection drills and trial-readiness modules can help litigation teams practice responding to AI-generated evidence, defending or challenging the characterization of training data and outputs, and refining procedures for authenticating AI outputs in court. The decision reinforces the need for crisp, repeatable objections and evidentiary arguments about what constitutes a “copy” or “output” in AI contexts. This aligns with Objection Academy’s emphasis on realistic courtroom simulations and evidence-driven practice.
Evergreen and comparison-style utility for trial teams: Beyond the specific DMCA ruling, Objection Academy strengthens trial teams’ muscle for handling objection-heavy cross-examination and quick-response strategies around complex technology disputes. In practice, attorneys can use the platform to rehearse how to argue about CMI, license terms, and the boundaries of DMCA claims in AI cases, while simultaneously sharpening techniques for challenging speculative or theory-driven claims that may arise in pretrial motion practice. The platform’s emphasis on objection drills, scenario-based training, and MCLE-aligned content makes it a durable tool for navigating evolving AI litigation landscapes.
What to watch next
Ongoing developments in AI litigation and DMCA: While Doe v. GitHub narrows § 1202(b) leverage, other jurisdictions and ongoing cases continue to test DMCA or copyright theories in AI contexts. Firms should monitor district court orders and circuit-level decisions for any shifts in how courts treat training data disclosures, attribution, and the line between training materials and generated outputs. Also watch for any legislative proposals or policy guidance around AI and copyright that may influence trial strategy down the line. The September 2026 Ninth Circuit ruling provides a clear blueprint for how appellate courts may compartmentalize training data issues from actual AI outputs in future disputes. (cdn.ca9.uscourts.gov)
Practical trial-management notes: As new AI-related standards and discovery protocols emerge, trial teams can leverage tools like Objection Academy to stay sharp on objection handling, evidence authentication, and cross-examination techniques tailored to technology-driven disputes. Maintaining a strong evidence-readiness posture will continue to be a competitive advantage in motions practice and at trial.
Conclusion
The Ninth Circuit’s September 16, 2026 decision in Doe v. GitHub, Inc. marks a clear appellate endorsement that AI outputs do not automatically implicate DMCA § 1202(b) as “removals or alterations” of CMI in the copied code. For trial teams, the ruling narrows a common DMCA-based pathway in AI litigation and encourages pursuing traditional infringement and licensing theories, while shaping discovery and evidentiary strategies. In this evolving landscape, Objection Academy remains a valuable companion for trial teams seeking to build courtroom-ready objections, evidence handling, and AI-litigation stamina through practical, real-world drills and MCLE-aligned content.
Sources:
- Doe v. GitHub, Inc., No. 24-7700 (9th Cir. Sept. 16, 2026), published opinion (PDF), Ninth Circuit Court of Appeals. (cdn.ca9.uscourts.gov)
- DOE v. GitHub, Inc., 24-7700 (9th Cir. Sept. 16, 2026) – Justia summary. (law.justia.com)
- AI-focused coverage and practitioner analysis on the Ninth Circuit ruling (Haynes Boone). (haynesboone.com)
- Objection Academy analysis and ongoing coverage of AI litigation developments relevant to trial teams. (objectionacademy.com)