Anyone who follows the tech industry knows that the lawsuits at this point are countless, however, a new entry filed this month against Microsoft-owned Github challenges the fundamental fundamentals behind some of the most advanced advancements. importance in artificial intelligence over the past three decades.
The lawsuit, led by programmer and attorney Matthew Butterick, focuses specifically on issues with Github Copilot, an AI assistant tool that offers programmers suggested code snippets as they code, much like the auto-complete feature in Google Docs or Gmail. Copilot learned what kinds of lines to code after fetching huge samples of publicly available lines of code from the internet. During this process, the proposed class action court case alleges that Copilot blatantly ignores or suppresses licenses presented by software engineers and effectively relies on “software piracy on an unprecedented scale”.
“It is not right, authorized or justified,” the lawsuit read. “Rather, 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 licenses chosen by open source programmers and monetizes their code despite GitHub’s commitment never to do so.
In a separate space blog post, Butterick says Microsoft’s approach with Copilot creates a “walled garden,” making it difficult for programmers in traditional open source communities. If this continues, he argues, it will starve open source communities and, over time, kill them.
Rather than accuse Microsoft and Github of violating copyright laws, Butterick’s lawsuit accuses Copilot of violating the companies’ terms of service and privacy laws and of breaching the Federal laws that require companies to display the copyright information of materials they use. And while this particular combination focuses on Copilot in particular, the tenets of the argument potentially apply to many, many other tools in place that use similar scraping methods to develop their tools.
“If companies like Microsoft, GitHub, and OpenAI choose not to follow the law, they shouldn’t expect us the public to stand still,” Butterick said in a recent blog post. “AI must be fair and ethical for everyone. If not, then it can never achieve its vaunted goals of uplifting humanity. It will simply become another means for a privileged few to benefit from the work of the greatest number.
“We have been committed to innovating responsibly with Copilot from the start and we will continue to evolve the product to best serve developers around the world,” a Github spokesperson said in an email to Gizmodo.
Microsoft did not respond to a request for comment.
“A better new world of software piracy”
These concerns about AI copyrights and compensation are not limited to programmers. Writers, the musiciansand visual artists have all echoed these concerns in recent years, especially in the wake of increasingly popular and effective generative AI image and video tools like Open AI’s DALL-E and Stable Diffusion. Unlike previous AI training which stuffs inelegantly billions of data units in a training set for an AI system, new generative approaches like DALL-E will take images of Pablo Picasso and then turn them into something new based on the description of a user. This act of repurposing data further complicates traditional copyright thinking. Like Butterick, a growing chorus of creative artists and writers has gone public recently express understandable fears the coming maturity of the AI system threatens to put them out of work.
Some companies are exploring new ways to credit people whose work ends up influencing the algorithm. Last month for example, Shutterstock announcement he would start selling DALL-E’s AI-generated art (also trained on humans) directly on his website. As part of the initiative, Shuttersock said it would launch an initial “contribution fund” to compensate contributors whose Shutterstock images were used to help develop the technology. Shutterstock said it was also interested in compensating contributors with royalties when DALL-E uses their creations.
Whether or not this plan works in practice, however, remains unclear and Shutterstock is only a relatively small company compared to Big Tech giants like Microsoft. Industry-wide, proposed standards for compensating creators for inadvertently training AI systems remain non-existent.
“Copilot basically tasks you with marking a 12-year-old’s homework over and over again,” Butterick wrote.
Speaking of her recent costume, Butterick acknowledged the novelty of the complaint and said it would likely be amended in the future. Although probably the first such a legal effort to strike at the root of AI training, the programmer and lawyer said he believes it is an important step in holding AI creators accountable in the future.
“This is the first leg of what will be a long journey,” Butterick said. “As far as we know, this is the first class action case in the United States challenging the formation and release of AI systems. It won’t be the last. AI systems are not exempt of the law.