OpenAI Faces Early Appeal in First AI Copyright Suit From Coders

OpenAI Inc. and Microsoft Corp.‘s GitHub will head to the country’s largest federal appeals court to resolve their first copyright lawsuit from open-source programmers who claim the companies’ AI coding tool Copilot violates a decades-old digital copyright law.

Judge Jon S. Tigar granted the programmers’ request for a mid-case turn to the US Court of Appeals for the Ninth Circuit, which must determine whether OpenAI’s copying of open-source code to train its AI model without proper attribution to the programmers could be a violation of the Digital Millennium Copyright Act.

The programmers first sued OpenAI and GitHub in 2022, alleging that Copilot trained on billions of lines of open-source code posted on the website GitHub without permission and routinely reproduced the copyrighted code in outputs. Copilot, a joint venture between OpenAI and GitHub, allows users to input the start of a line of code and the AI produces suggestions to finish the remaining lines of code.

The case is the first of many against OpenAI claiming that training AI models on valuable intellectual property violates the law. The company is facing copyright infringement suits from authors and media outlets including The New York Times.

Tigar, who sits for the US District Court for the Northern District of California, had largely dismissed the programmers’ case in June, ruling that they failed to show that the AI coding tool could produce identical matches of the copyrighted code.

The judge’s Sept. 27 discretionary order sending the case to the Ninth Circuit said the issue of whether the programmers needed to show an identical match was a “controlling question of law” that goes to the core of the case and will resolve divided opinions among district courts.

The DMCA, which Congress passed in 1998 to update copyright law for the internet age, prohibits the removal of “copyright management information” that is attached to a copyrighted work. CMI can include authorship information and licensing terms.

The programmers argued that Copilot fails to include authorship and licensing terms when it outputs code. Unlike other lawsuits against AI companies, the programmers didn’t allege that OpenAI and GitHub engaged in copyright infringement, which is different from a DMCA violation.

Tigar’s Sept. 27 order also granted the programmers’ request to pause the proceedings while the Ninth Circuit resolved the appeal.

The Joseph Saveri Law Firm LLP and Matthew Butterick represent the programmers. Morrison & Foerster LLP represents OpenAI. Orrick Herrington & Sutcliffe LLP represents GitHub.

The case is Doe 1 v. GitHub Inc., N.D. Cal., No. 4:22-cv-06823, 9/27/24.