Microsoft and GitHub ask court to scrap lawsuit over AI-powered CoPilot – Courthouse News Service

Microsoft and GitHub ask court to scrap lawsuit over AI-powered CoPilot – Courthouse News Service

Updates to our Terms of Use
We are updating our Terms of Use. Please carefully review the updated Terms before proceeding to our website.
The plaintiffs claim Microsoft, GitHub and OpenAI are violating the coders’ term of use by not crediting them for the code the AI learns and copies.
(CN) — Attorneys for Microsoft, GitHub and the artificial intelligence research laboratory OpenAI asked a federal judge to dismiss a class action filed on behalf of four unnamed coders over the paid service CoPilot, which uses AI to suggest code to its users.
Owned by Microsoft, GitHub is a site that hosts open-source computer code. In much the same way that Creative Commons allows photos to be shared but only under certain conditions — by, say, crediting the photographer — code on GitHub typically comes with licensing agreements, often mandating that credit be given to the author of the code.
In November 2022, a handful of unnamed coders sued claiming GitHub’s CoPilot, which charges users $100 per year, violates the open-source licensing by not crediting the authors of the source code the AI has learned from. The complaint puts forward a number of causes of action, including breach of contract for the open-source license violations, fraud, unjust enrichment and unfair competition.
“CoPilot ignores, violates, and removes the licenses offered by thousands — possibly millions — of software developers, thereby accomplishing software piracy on an unprecedented scale,” the plaintiffs say in the complaint. “CoPilot’s goal is to replace a huge swath of open source by taking it and keeping it inside a GitHub-controlled paywall. It violates the licenses that open-source programmers chose and monetizes their code despite GitHub’s pledge never to do so.”
In their motion to dismiss the lawsuit, Microsoft and GitHub argued just the opposite: that CoPilot is very much in keeping with the spirit of open-source information sharing.
“CoPilot withdraws nothing from the body of open source code available to the public, the motion said. “Rather, CoPilot helps developers write code by generating suggestions based on what it has learned from the entire body of knowledge gleaned from public code. In so doing, CoPilot advances the very values of learning, understanding, and collaboration that animate the open source ethic. With their demand for an injunction and a multibillion-dollar windfall in connection with software that they willingly share as open source, it is plaintiffs who seek to undermine those open source principles and to stop significant advancements in collaboration and progress.”
During oral arguments, conducted Thursday over Zoom, Open AI attorney Joe Gratz objected to the idea that CoPilot copies any one individual piece of code.
“The code that is suggested is code that is common to a particular issue,” said Gratz. “It’s seen that thing many times.”
CoPilot relies on code that’s been seen many, many times. How is that supposed to be credited?
Matthew Broderick, one of the plaintiff’s attorneys, is also a coder and self-described “open-source programmer” and “member of the open-source community.” He sought to reframe the issue.
“When the AI trains, it’s not just reading,” he told U.S. District Judge Jon Tigar. “It’s copying. What’s being imprinted on the neural network is copying. When you, as a user, query it, it goes through all that data.”
But Tigar said he found it difficult to see how the AI software’s training — that is, scanning through all the code voluntarily uploaded onto to GitHub — constituted a violation.
“Does the act of copying it onto the neural network violate any of the GitHub licenses?” Tigar asked.
“Yes,” Broderick said. “The suggested codes’ licenses all contain attribution requirements.”
The Judge indicated that he was likely to dismiss at least some of the claims, especially the ones that alleged a “concrete particularized injury,” like the negligence claim, since the complaint never points to any specific instance of a plaintiff’s pieces of code being copied. But, he suggested, the plaintiffs might still be entitled to an injunction to prevent any future violation of a licensing agreement.
“It is a mathematical near-certainty that this product will eventually use material from class members,” the judge said, taking the matter under submission.
It’s been quite a year for OpenAI, which among other things owns ChatGPT, the fourth iteration of which was released this year. The chatbot was widely seen as a quantum leap in the development of AI, and has sparked fears that the technology is progressing faster than humans are able to grapple with.
In an oft-cited survey taken in 2022, AI experts were asked, “What probability do you put on human inability to control future advanced AI systems causing human extinction or similarly permanent and severe disempowerment of the human species?” The median answer was 10%.
This lawsuit, though, is among the first to test the legal boundaries of how the current crop of artificial intelligence works — by scraping available information and creating a sort of composite out of material it judges to be reliable or commonly used. Should the suit prove successful, it could put the brakes on the development of such software — just what many of its critics want, if for different reasons.
Sign up for new weekly newsletter Closing Arguments to get the latest about ongoing trials, major litigation and hot cases and rulings in courthouses around the U.S. and the world.

source