Date of Award


Document Type


Degree Name

Doctor of Juridical Science (SJD)


The two groups of the human-or-machine questions, whether AI-generated works are copyrightable and whether AI-generated works have human authors, are revisiting the current copyright law with the emergence of AI-generated works. These revisiting questions reveal that the current authorship requirement fails to provide a clear and operable standard on evaluating a human contributor’s intellectual labor for creative output. Such a defect of the current authorship requirement has to be fixed to respond to the technological change of artificial intelligence and the burgeoning prevalence of AI- or advanced computer program-generated works.

This dissertation’s main goal is to fix the flaw of the authorship requirement by establishing an improved authorship spectrum. The improved authorship spectrum can serve as a guide to evaluate whether a human contributor provide sufficient intellectual labor for creative output, and to locate the human author(s) behind a creative output in this AI era. I argue that by applying the improved spectrum to AI-generated works, such types of works can be distinguished into the two categories “the authored and copyrightable AI-generated works” and “the authorless and uncopyrighted ones.” Therefore, my intended conclusion for the revisiting human-or-machine questions is: not every AI-generated work falls out of the scope of copyright protection; some of the AI-generated works do have human authors and thus are copyrighted works of authorship, but some are authorless works because their human contributors all failed to offer the sufficient intellectual labor for the work.