GitHub, Microsoft and OpenAI will have to defend their usage of third-party programming code in a U.S. court. The three parties wanted to get rid of two specific allegations before the case actually starts. For example, they will still have to argue the legitimacy of Codex reproducing code that is “borrowed” from others. In addition, the lawsuit will still determine whether Copilot and Codex share enough authorship information when displaying copyrighted code. The judge wants these claims addressed in court, much to the chagrin of the three parties.
In November, developers sued the three parties for software piracy. Early this year, GitHub, parent company Microsoft and OpenAI let it be known that they disagreed with two specific allegations in this area. Thus, they wanted the lawsuit to not be about these either.
There are quite a few privacy and copyright concerns surrounding the deployment of AI tools. AI models such as OpenAI’s GPT-3 and GPT-4 rely on massive datasets, many containing open-source code and Web content.
The Register now reports that Judge Jon Tigar does see something to the charges. He ruled with prejudice, which under the U.S. legal system prevents the defence from appealing the decision. According to the judge, the defence’s complaints were not detailed enough to proceed.
It is not yet entirely clear what the final charges will look like or when we can expect a trial. What is certain is that the lawsuit has a lot at stake for Microsoft and Co. If the use of copyrighted code is not legitimately displayed in Copilot and Codex, in all likelihood, it will have to be modified in a ChatGPT window as well. In addition, political concerns about privacy in AI and copyright within AI datasets will have a potential deterrent effect on the integration of artificial intelligence into commercial products, such as Office 365.