As universities adapt to the advancements of artificial intelligence, legal restraints are threatening the possible return to traditional research and writing methods. The outcome could impact access to AI tools at universities across the U.S., tools that have become a major asset to students in the past year.
Students at the university have their own reasons for incorporating AI in their academic studies. The potential disappearance of ChatGPT may cause distress among the students if they are forced to readjust their habits.
A computer science junior, Sydney Divozzo said, “I use AI to help me get an idea of how I should code my projects for my classes. When coding in Python, it is helpful to use AI to create an outline for the project that I am working on. If ChatGPT were to leave, I would be really pissed.”
The New York Times filed a lawsuit against OpenAI and Microsoft, alleging copyright violation on the grounds that these companies utilized millions of their articles to train AI technologies, including ChatGPT on Dec. 27.
On Feb. 16, California Judge, Araceli Martinez-Olguin, dismissed claims against Open AI by two best-selling authors, Sarah Silverman and Paul Tremblay, for illegally using their novels to train AI language tool, ChatGPT, and create human-like responses. The case was ultimately dismissed because the authors couldn’t prove a direct similarity between ChatGPT responses and their books.
These legal battles reflect the increasing complexity of intellectual property battles and the rapid advancements in AI.
“Producing original independent journalism is at the heart of this mission. The Times’s newsroom produces groundbreaking journalism across every major storytelling format,” The New York Times stated in their complaint to the U.S. District Court for the Southern District of New York.
Michael Froomkin, a law professor at the University of Miami, highlights the critical role web scraping–the process of extracting data from web pages using machine learning techniques–plays in this legal battle. As technology rapidly advances, legal frameworks and guidelines regarding the utilization of web scraping technologies are still in the process of being established.
“Scraping is the most important question. What will AI be capable of, scraping the whole internet and training data with copyright content?” Froomkin said.
AI companies have claimed in an OpenAI blog post that they are able to lawfully use such content obtained by scraping for training their technologies since the material is public, and they are not duplicating the content verbatim.
In their complaint, The New York Times calls for the destruction of all GPT or similar LLM models and the training datasets that incorporate materials from The New York Times. This request aims to prevent further unauthorized use of their copyrighted works.
According to a survey conducted by BestColleges, 54% of college students utilize AI tools in their studies, while 53% have encountered assignments that require the use of AI.
Froomkin pointed out that content generated by AI isn’t technically free of copyright, noting that, “students need to be sensitive to copyright and just because it came from AI it doesn’t mean it’s okay.”
The courts will ultimately decide whether these AI technologies illegally copied content from sites like The New York Times which can affect the legal definition of copyright in the context of AI-generated content.
If The Times wins, the court may order AI tech companies to destroy their data sets and start over using original license work. However, it is unknown whether that will be possible and how long the process would take.
The legal battles against OpenAI mark a major change in how content is made and used due to the potential impact on intellectual property rights and the regulations of AI technologies. The outcomes of these cases may shape not only current practices but future legislative initiatives.