Appeals Court Ends the Copilot Copyright Case

The long-running copyright fight over GitHub Copilot has reached its end at the appeals stage. On 16 September a three-judge panel of the US Court of Appeals for the Ninth Circuit rejected the central claim brought by a group of anonymous programmers against GitHub, its owner Microsoft and OpenAI, upholding a lower court's dismissal.

The programmers argued that Copilot, the AI coding assistant trained on public repositories, stripped away the copyright notices and author names attached to their code when it suggested similar snippets to other developers. They pinned that argument on Section 1202 of the Digital Millennium Copyright Act, which bars the intentional removal of copyright management information.

The panel was not persuaded. As the Electronic Frontier Foundation summarised the reasoning, the absence of copyright information in a newly generated work does not by itself prove that someone illegally removed it. Section 1202, the court held, was meant as "a backstop for traditional copyrights in the digital age", not a new and broader right.

Two doors remain open, though neither revives the case as filed. Programmers can still bring ordinary copyright infringement claims if a model reproduces their code outright, and the plaintiffs' separate contract-based claims survive.

The first case of its kind

This was the first copyright suit filed against OpenAI and its generative products, and for years it was watched as a bellwether for how courts would treat models trained on public material. Bloomberg Law reported the ruling as a clean win for the companies.

It lands in a busy season for AI and the law. The Justice Department recently said it would pursue AI harms through prosecution rather than new rules, and a separate settlement in Bartz v. Anthropic put a price on training with pirated books. The Copilot decision narrows one avenue, DMCA removal claims, without settling the harder question of what counts as infringing output.