Microsoft, GitHub, and OpenAI have petitioned a U.S. court to dismiss a lawsuit that accuses them of violating open-source licenses and seeks $9 billion in compensation for GitHub Copilot. Lawyers for Microsoft, GitHub, and OpenAI argue that the allegations of license infringement and claims for compensation are unfounded.
The companies argue that there is no evidence of damage in the lawsuit and no substantiation for the claims, including examples of violations of legal rights. Microsoft, GitHub, and OpenAI allege that the plaintiffs rely only on “hypothetical events, and GitHub Copilot did not personally harm them.
Lawyers for the companies argue that the lawsuit does not specify specific copyrights that GitHub Copilot violated through unauthorized use, or contracts and licenses violated in the creation of the artificial intelligence tool.
Microsoft says the allegations of copyright infringement are unfounded and contrary to the doctrine of fair use, which allows copyrighted material to be used in certain situations. Microsoft and GitHub cite a 2021 U.S. Supreme Court decision that confirms that Google’s use of Oracle source code to build the Android operating system is a good faith use.
Copilot will not degrade the Open-Source community
According to Microsoft and GitHub, Copilot does not remove anything from publicly available open source, but only helps developers write code by offering options based on the information that the tool has received from public code.
IT companies, through their lawyers, said in response to the court that the plaintiffs were undermining open-source principles by demanding an injunction and billions in compensation for open-source software. A lawsuit against GitHub Copilot is scheduled for May.
Recall that in early November 2022, Matthew Butterick, a programmer-lawyer, filed a lawsuit in California against Microsoft, GitHub, and OpenAI of the year because of the violation of the licenses of open source projects and infringement of the rights of programmers to the neural network assistant GitHub Copilot. The developer is demanding $9 billion in compensation from U.S. companies.
Copilot’s GitHub currently uses millions of lines of code from public GitHub repositories to generate code and can translate natural language into code in dozens of programming languages. The neurofeedback performs this work automatically, without analyzing or considering the licensing rules of open source projects, such as the GPL, Apache, or MIT, which require attribution and specific copyright definition when using the code.
Copilot removes links to open-source licenses
Microsoft, GitHub, and OpenAI took precautions in developing Copilot. The tool has been configured to remove any link to open-source licenses from the generated code, even if it copies snippet code longer than 150 characters from a particular repository. The system does not indicate the authorship of the source code fragment.
Developer Butteric suggests that every time Copilot produces a virtually illegal result, this neural network tool violates the DMCA law three times by distributing materials with a license without attribution, copyright notice, and license terms. The developer believes that in 15 months of work, Copilot, Microsoft, GitHub, and OpenAI about 3.6 million times violated US laws and caused damage to the open source community in the amount of 9 billion US dollars.
Batteric worries that Copilot could further degrade the open source community, and as a consequence, the quality of the code used in the training data of a system that is unable to generate code on its own will decline.