GitHub faces lawsuit over Copilot AI coding assistant

Class-action complaint contends that training the AI system on public GitHub repos violates the legal rights of creators who posted the code under open-source licenses.

Thinkstock

While GitHub continues to trumpet the effectiveness of GitHub Copilot, its AI-driven coding assistant, a class-action lawsuit has been filed that challenges the legality of the technology. The lawsuit claims that GitHub’s training the Copilot AI on public GitHub repositories has violated the rights of the “vast number of creators” who posted code under open-source licenses on GitHub.

Filed in the US District Court in San Francisco on behalf of perhaps millions of GitHub users, the class-action complaint alleges that “Copilot ignores, violates, and removes the licenses offered by thousands—possibly millions—of software developers, thereby accomplishing software piracy on an unprecedented scale.” The litigation seeks recovery of damages and restitution.

The lawsuit is believed to be the first class-action case in the US chal­leng­ing the training and out­put of AI sys­tems. “It will not be the last,” said lawyer and programmer Matthew Butterick, who spearheaded the lawsuit, in a bulletin on the action. Besides GitHub, defendants include GitHub owner Microsoft and OpenAI, the co-developer of Copilot.

GitHub Copilot is tool that suggests code snippets and functions in real time, right from the developer’s code editor. It has been trained on billions of lines of code. “Spend less time creating boilerplate and repetitive code patterns, and more time on what matters: building great software,” the GitHub Copilot website advises. But the tool immediately drew the ire of the Free Software Foundation, which cried foul on Copilot’s use of freely licensed software.

GitHub, in response to the lawsuit, released a statement this week defending the technology while pledging to evolve. “We’ve been committed to innovating responsibly with Copilot from the start, and will continue to evolve the product to best serve developers across the globe.”

The next major release of Copilot is slated to provide the ability to identify strings that match public code with a reference to those repositories. With this information, developers might be inspired by other codebases and could gain confidence that a fragment is appropriate for use in their project, GitHub noted. GitHub said businesses soon will be able to purchase and manage Copilot seat licenses for employees.