At every stage of the process, from instructing data to inquiries about the controller of the outcome, copyright queries present a challenge to relationalAI.
Adi Robertson, a former technology and policy editor, specialized in free speech, online platforms, and virtual reality. Since 2011, Adi has contributed to The Verge covering topics such as video games, biohacking, and more.
In 2022, James Allen filed a trademark application for an image titled “Theatre D’opéra Spatial,” or “Space Opera Theater.” Despite the image bearing a striking resemblance to a detailed sci-fi painting, Allen meticulously experimented with the AI image generator Midjourney to produce it. He mentioned undergoing over 600 rapid iterations to achieve the desired outcome. However, the US Copyright Office did not view Allen’s work favorably. His copyright request became one of the initial rejections specifically related to the use of AI tools, highlighting the persistent copyright challenges posed by such technologies.
Fundamental questions regarding artists’ rights over their creations have been brought to light by conceptual AI programs, particularly image generators like Midjourney, Stable Diffusion, and DALL-E. In the United States, several legal battles may determine whether these programs qualify for the fair use exceptions, given that many are trained on substantial datasets containing copyrighted material without artists’ consent. Individuals who perceive DALL-E or Stable Diffusion as akin to tools like Photoshop are frustrated by the US Copyright Office’s stance that a computer program cannot generate copyright-protected art.
While copyright law recognizes that most art builds upon preceding works, exceptions like fair use exist. Copyright law aims to foster artistic creation. When a new technology like AI emerges, it may necessitate a reassessment of the delicate balance. In the realm of AI, where will the scales tip? The future remains uncertain.
