Generating copyright law in the world of artificial intelligence
Presentation Type
Abstract
Faculty Advisor
Thomas Loikith
Access Type
Event
Start Date
25-4-2025 9:00 AM
End Date
25-4-2025 9:59 AM
Description
The rise of foundation models like Open AI’s ChatGPT raises important questions about whether training on copyrighted materials constitutes infringement under US copyright law. Recent lawsuits illustrate the growing tensions between copyright holders and artificial intelligence (AI) developers. Should copyright law be amended to address the unique challenges that AI presents? Specifically, does the “fair use doctrine” apply to disputes between copyright holders and AI developers? How should concerns such as the lack of dataset transparency and potential market effects be addressed? What should be done to balance intellectual property protection with rapid advancements in AI technology? As the use of AI expands, this is a significant question of interest to not only copyright holders and AI developers, but also to society in general. This is a complex question requiring interdisciplinary research and analysis. Using qualitative research methodology (textual analysis), I analyzed sources from the disciplines of law and computer science to determine conflicting insights. I then found common ground between the insights and integrated them to arrive at a more comprehensive understanding of the research problem. That more comprehensive understanding leads to suggested resolutions of the problem. I anticipate completing my research in early April. I can conclude, however, that the transformative nature of GenAI outputs and current safeguards against direct copying support claims of “fair use” by AI developers. Further, lack of dataset transparency and potential market effects can be addressed by amending copyright law to require transparency in training datasets and implementing opt-out programs for copyright holders.
Generating copyright law in the world of artificial intelligence
The rise of foundation models like Open AI’s ChatGPT raises important questions about whether training on copyrighted materials constitutes infringement under US copyright law. Recent lawsuits illustrate the growing tensions between copyright holders and artificial intelligence (AI) developers. Should copyright law be amended to address the unique challenges that AI presents? Specifically, does the “fair use doctrine” apply to disputes between copyright holders and AI developers? How should concerns such as the lack of dataset transparency and potential market effects be addressed? What should be done to balance intellectual property protection with rapid advancements in AI technology? As the use of AI expands, this is a significant question of interest to not only copyright holders and AI developers, but also to society in general. This is a complex question requiring interdisciplinary research and analysis. Using qualitative research methodology (textual analysis), I analyzed sources from the disciplines of law and computer science to determine conflicting insights. I then found common ground between the insights and integrated them to arrive at a more comprehensive understanding of the research problem. That more comprehensive understanding leads to suggested resolutions of the problem. I anticipate completing my research in early April. I can conclude, however, that the transformative nature of GenAI outputs and current safeguards against direct copying support claims of “fair use” by AI developers. Further, lack of dataset transparency and potential market effects can be addressed by amending copyright law to require transparency in training datasets and implementing opt-out programs for copyright holders.
Comments
Poster presentation at the 2025 Student Research Symposium.