Judge Tigar allows only two of the original 22 claims to proceed, signaling a significant shift in the way copyright law is interpreted in AI-generated content.
Microsoft, its subsidiary GitHub and business partner OpenAI experienced a significant win after most of the claims of a lawsuit brought against them were dismissed by a California court. The legal battle, which started in November 2022 when programmer and lawyer Matthew Butterick filed suit, will proceed on only two out of the 22 initial claims.
California Northern District Judge Jon S. Tigar ruled that the plaintiffs’ claim GitHub Copilot violates section 1202(b) of the Digital Millennium Copyright Act does not hold merit. The judge determined that the code snippets provided by GitHub Copilot to users are not similar enough to copyright-protected codes created by the plaintiffs.
Lawsuit Backgrounds
In an initial statement, plaintiff Matthew Butterick claimed, “This is the first class-action case in the U.S. challenging the training and output of AI-systems.” Other plaintiffs joined the action, claiming OpenAI has scrapped code pieces published by programmers on GitHub collaboration platform (owned by Microsoft).
GitHub Copilot launched in June 2021. Its purpose is to suggest lines of codes to programmers directly into their code editor. The AI code-generator was described in the press as “capable of suggesting the next line of code as developers type.”
According to the Doe(s) v. GitHub Inc., et al. lawsuit, these suggestions reproduce line-by-line copyright protected code. They are seeking $1B in damages. In January 2023, Microsoft, GitHub and OpenAI petitioned the court to dismiss the lawsuit. However, their request was denied.
The Groundbreaking Court Ruling
United States District Judge Jon S. Tigar issued a 15-page order, dismissing the most important claims in the lawsuit. Probably the biggest blow to the plaintiffs’ case is the ruling out of the claim that the AI-powered GitHub Copilot violates DMCA 1202(b).
This clause refers to removing without permission essential “copyright management” details, including the name of the code writer and the terms of the license. However, Judge Tigar decided that clause 1202(b) cannot be invoked in this case. The reason for it is that the code generated by GitHub Copilot is not similar enough to infer copyright infringement.
The judge also dismissed the core claim that GitHub Copilot is capable of, and actually generating code that is identical to their own work. Judge Tigar argued that the plaintiffs did not produce a single piece of evidence demonstrating this claim. The plaintiffs still have the right to amend this claim.
The judge also dismissed an “unpersuasive” study cited by the plaintiffs as evidence that generative AI models increase their capacity of memorizing training data and offering it as an exact duplicate in its output.
Finally, Judge Jon Tigar dismissed the claims for punitive damages and for monetary relief for unjust enrichment.
Implications of the Ruling for the Future of AI
Although the plaintiffs’ lawsuit continues on two counts of breach of contract and open-source license violation, the most important claims were ruled out. The court’s stance hails a shift in the treatment of AI-generated content from a copyright standpoint.
What started as “the Napster-era of AI” from the programmers’ point of view in 2022, has now reached a turning point. Judge Tigar’s ruling may be interpreted as a green light to create more generative AI models by other companies.
Microsoft and OpenAI are currently involved in other legal actions related to copyright infringement by AI-generated content. There is no telling whether Judge Tigar’s decision will have any impact on the rulings in these cases.
One thing is clear: The distribution of supporters and detractors is fairly even and no party appears to have the stronger argument. For legal professionals, this is a huge opportunity to shape future laws and to bring a level of certainty and regulation to a still unexplored legal landscape.