Data Scraping for Generative AI—To What Extent?

Generative AI platforms are developed by scraping data from other platforms. This raises complex copyright infringement issues that are yet to be resolved. Courts have attempted to address these problems using the fair use doctrine, which evaluates four factors to determine whether a use infringes on copyright: (1) the purpose and character of the use, (2) the nature of the use, (3) the amount and substantiality of the copyrighted work’s use, and (4) the impact of the use on the market for the original work. Despite efforts to apply this long-established doctrine, challenges persist, including difficulties in identifying the copyright owner, distinguishing between inputs and outputs, and assessing transformativeness. As a result, courts have suggested handling these cases on an ad-hoc basis and through jury trials, leaving many questions unresolved. This Note examines the limitations of the fair use doctrine in relation to generative AI and proposes additional factors for consideration in resolving these issues.