The GitHub Copilot lawsuit, filed in November 2022 as Doe v. GitHub, Inc. in the U.S. District Court for the Northern District of California, has been narrowed to a single breach-of-contract claim and is currently stayed while the Ninth Circuit decides whether the Digital Millennium Copyright Act protects author attribution only on exact copies of a work. The plaintiffs, a proposed class of open-source developers, accuse GitHub, Microsoft, and OpenAI of training Copilot on publicly posted code without honoring the attribution terms of licenses such as MIT, GPL, and Apache. Penalties sought exceed $9 billion.1Courthouse News Service. Judge Trims Code-Scraping Suit Against Microsoft, GitHub
What the Lawsuit Alleges
GitHub Copilot is a paid AI assistant that suggests code to developers as they type. It runs on OpenAI’s Codex model, which was trained on huge volumes of publicly hosted source code — 159 gigabytes of Python alone, drawn from more than 54 million repositories in the original 2021 version, along with JavaScript, Go, Ruby, C++, and other languages.
Much of that code carried open-source licenses that permit free reuse only if the redistributor keeps the original author’s name, copyright notice, and license text attached. The complaint says Copilot strips all of that away and hands the code to paying subscribers stripped clean. The suit was brought by anonymous plaintiffs identified as “J. Doe 1” and others, represented by attorney Matthew Butterick and the Joseph Saveri Law Firm. The proposed class covers developers who published code under any of 11 attribution-requiring open-source licenses.2GitHub Copilot Litigation. GitHub Copilot Investigation
The complaint originally packed in more than 20 claims. Two theories carried the case:
- Breach of the open-source licenses. The plaintiffs treat each license as a binding contract. When Copilot outputs code without the author’s name, notice, or license terms, it breaks that contract.3Finnegan. Insights From the Pending Copilot Class Action Lawsuit
- DMCA Section 1202(b). That provision bars the intentional removal of “copyright management information” — author names, notices, license text — when the person doing the removing knows it will facilitate infringement. The plaintiffs say Copilot does exactly that at training time and again at output.2GitHub Copilot Litigation. GitHub Copilot Investigation
What the District Court Has Already Decided
Judge Jon S. Tigar has ruled on two rounds of motions to dismiss, and each round narrowed the case.
In May 2023, he dismissed claims for civil conspiracy and declaratory relief with prejudice, letting the license and DMCA theories move forward.1Courthouse News Service. Judge Trims Code-Scraping Suit Against Microsoft, GitHub
The larger cut came on January 22, 2024. Judge Tigar threw out state-law claims for tortious interference, unjust enrichment, negligence, and unfair competition, ruling that they were preempted by the Copyright Act because each ultimately turned on unauthorized reproduction of code. Two plaintiffs also lost their damages claims for lack of standing, though three others could keep seeking money.1Courthouse News Service. Judge Trims Code-Scraping Suit Against Microsoft, GitHub
The heart of that order dismissed the DMCA claim. The judge held that Section 1202(b) reaches only situations where copyright management information is removed from an identical copy of a protected work. The plaintiffs’ own complaint had described Copilot’s output using words like “modification,” “near-identical copy,” and “functional equivalent” — language the court said undercut any claim that Copilot was distributing exact reproductions. He gave the plaintiffs leave to amend but wrote that he found it “unlikely that this deficiency could be cured by the allegation of additional facts,” granting the chance only “out of abundance of caution.”4ChatGPT Is Eating the World. Judge Tigar Dismisses DMCA CMI Claims v. GitHub
That leaves one claim standing at the trial-court level: breach of contract based on the open-source license terms. Judge Tigar rejected OpenAI’s argument that the plaintiffs should have brought a copyright-infringement suit instead, holding that plaintiffs may pick their own theory.5LWN.net. Judge Tigar’s Rulings on the GitHub Copilot Lawsuit
The Ninth Circuit Appeal
Rather than wait until the whole case ends, the plaintiffs asked the judge to certify the DMCA dismissal for immediate appeal. He agreed. The appeal, docketed as No. 24-6136 in the Ninth Circuit, presents one question: whether Sections 1202(b)(1) and (b)(3) contain an “identicality requirement,” or whether the statute also applies when copyright management information is stripped from modified or derivative versions of a work.6Venable. DMCA Question Certified for Appellate Court
The district court has stayed everything else, including the surviving breach-of-contract claim, until the appeal is resolved.7NYU Journal of International Law and Politics. Annotation on Doe v. GitHub
The plaintiffs filed their opening brief on April 9, 2025, arguing that an exact-copy rule would leave the DMCA toothless against AI systems that alter works slightly before redistributing them.8GitHub Copilot Litigation. Case Updates ACT | The App Association filed an amicus brief backing that position, warning that a strict identicality rule would leave open-source developers with no remedy when AI tools remove their attribution.9ACT | The App Association. Amicus Brief in Doe v. GitHub
The Chamber of Progress and the Computer and Communications Industry Association filed on the other side in July 2025. They argued that dropping the identicality requirement would turn the DMCA into a “dangerous super-copyright” that lets plaintiffs sidestep fair use and registration defenses while collecting statutory damages of $2,500 to $25,000 per violation.10CCIA. Amicus Brief of Chamber of Progress and CCIA in Doe v. GitHub
The Ninth Circuit heard oral argument on February 11, 2026. As of mid-2026, no decision had issued.11Baker McKenzie. The Copilot Litigation
Why the Ruling Matters Beyond Copilot
District courts have already split on the identicality question. Courts in the Northern and Central Districts of California have applied it; courts in the Southern District of Texas and Nevada have rejected it.6Venable. DMCA Question Certified for Appellate Court
The reasoning has already spread. In Andersen v. Stability AI, Judge William Orrick dismissed DMCA claims against the makers of Stable Diffusion in August 2024, explicitly following Judge Tigar’s analysis and holding that the plaintiffs had not alleged any AI-generated image was identical to a specific plaintiff’s work.12Justia. Andersen v. Stability AI Ltd. Two New York judges have gone opposite directions in cases against OpenAI over ChatGPT: Judge Colleen McMahon dismissed Raw Story Media’s DMCA claim in November 2024 for lack of standing, while Judge Jed Rakoff let The Intercept’s parallel claim proceed based on specific verbatim reproductions.13Skadden. Recent Decisions on Whether AI Training Violates the Digital Millennium Copyright Act
If the Ninth Circuit rejects the identicality requirement, AI developers whose models produce modified versions of copyrighted works face a new front of DMCA exposure. If it affirms, plaintiffs across the circuit will generally need to show exact reproduction before Section 1202(b) applies.
What GitHub and Microsoft Have Done
GitHub has said verbatim reproduction is rare, citing a figure of roughly 0.1% of suggestions. The Free Software Foundation has noted that this figure does not count near-verbatim copying with only variable-name changes.14Free Software Foundation. Copyright Implications of the Use of Code Repositories to Train a Machine Learning Model
The company has since added a duplicate-detection filter that blocks suggestions closely matching public code, though it does not yet cover all Copilot features. In September 2023, Microsoft added an IP indemnification policy under which it will defend paying users sued over unmodified Copilot suggestions. GitHub has also stated that it does not use data from Copilot Business or Enterprise customers for model training, and organizations can configure content exclusions to keep sensitive repositories out of Copilot’s indexing.15Microsoft Tech Community. Demystifying GitHub Copilot Security Controls
The defendants have previewed a fair use defense: if training on public code is fair use, no license was needed and the breach-of-contract theory collapses too. The court has not reached that argument because the case has not moved past the motion-to-dismiss stage.3Finnegan. Insights From the Pending Copilot Class Action Lawsuit
Where Things Stand and What to Watch
The case remains active in the Northern District of California before Judge Tigar, with the most recent filing recorded on May 6, 2026.16CourtListener. Doe 1 v. GitHub, Inc. All district court proceedings, including the surviving breach-of-contract claim, are paused pending the Ninth Circuit’s decision. No class has been certified, and no discovery schedule or trial date has been set.
The next milestone is the Ninth Circuit’s ruling on identicality. That decision will determine whether the DMCA claim comes back to life and, more broadly, how the statute applies to generative AI systems across the circuit. If the DMCA claim revives, the case returns to Judge Tigar’s docket alongside the breach-of-contract theory; if the dismissal is affirmed, the contract claim moves forward on its own.