ChatGPT and allegations of IP theft

Every once in a while, there emerges an innovation so useful that it affects how people go about their lives. We witnessed this with the emergence of Google, Facebook, and the iPhone. OpenAI’s newly launched service, ChatGPT, could have the same lasting impact.

Artificial intelligence. ChatGPT is a chatbot that uses artificial intelligence (AI) to provide detailed responses to users’ instructions. It is not the first invention to make use of AI. Years before ChatGPT was launched, people were already enjoying the services of virtual assistants such as Apple’s Siri and Amazon’s Alexa. In fact, even before Siri and Alexa, people were already relying on search engines like Yahoo and Google.

What makes ChatGPT unique, however, is its ability to accomplish complex tasks. It goes beyond answering trivial questions. The app can write poems, design a game, and solve mathematical equations—all in a matter of seconds. Because of its ability to accomplish these tasks quickly, efficiently, and effectively, ChatGPT has earned praise from many groups including programmers and business owners.

As with many technological advances, however, ChatGPT comes with a downside, in this case, the claim that it enables plagiarism. Let’s say a student needs an essay for class. He types, “draft a one-page essay on World War II” into ChatGPT, which then produces the essay. If the student submits the essay as his own, is he liable for plagiarism?

OpenAI’s terms of use actually assign to the user all the rights to its output. This means that the user owns what is generated by the app. Therefore, a student who generates the essay, strictly speaking, owns it. However, plagiarism has been defined as “the deliberate and knowing presentation of another person’s original ideas or creative expressions as one’s own.” Since the substance of the essay did not originate from the student, his act can still be construed as plagiarism.

ChatGPT does not generate responses from scratch. The model is trained to give responses based on data collected by its developer. Notably, it does not provide attributions for such sources. In the example, if ChatGPT produces an output in the form of an essay based largely on the work of a renowned historian, and the student submits the essay as his own, it can be argued that the student is effectively plagiarizing the historian’s work.

Copyright infringement. ChatGPT has also been accused of committing copyright infringement. Copyright is the legal protection given to people for their original work which may consist of literary, scientific, or artistic creations. Under our jurisdiction, such works are protected from the moment of creation, and a person who uses another’s work without consent commits copyright infringement.

Recently, a class-action lawsuit was reportedly filed against OpenAI. In the suit, authors allege that their copyrighted work was used to train ChatGPT without their consent. If the allegations are found to be true, they could subject not only OpenAI but also users who made use of such copyrighted work to potential liability.

Looking forward. It bears emphasizing that OpenAI is exerting efforts to address these legal issues. For one, it is reportedly developing a tool that can identify texts coming from ChatGPT, and eventually help educators detect potential cases of plagiarism. OpenAI’s website also allows creators to report users who are infringing upon their copyright. Users who are adjudged guilty of committing such acts could have their accounts terminated. Nevertheless, these mechanisms are by no means foolproof solutions. As such, for now, users will have to tread carefully when using this technology.

Mario C. Cerilles Jr. is cofounder of Cerilles and Fernan Intellectual Property Law. He also teaches intellectual property law at the University of the Philippines College of Law. Email: mccerilles@gmail.com