Artificial intelligence (“AI”) technology has detonated an explosive burst of seemingly creative expression. Stories, images, music, and even entire books are now being generated very quickly. This development is a major headache for copyright registrars because the copyrightability of works created in this way is uncertain. The almost limitless variability in the extent of human involvement in the creation of a work using an AI tool compounds the uncertainty. In some cases, copyrightability is easy to determine, such as where an author only claims rights in the selection and arrangement of AI-generated output rather than the output itself. But in many cases the registrability of a work created with the aid of an AI technology is far from certain. In these situations, the Copyright Office should apply a rule of doubt to allow registration. This solution appears to be a novel idea, at least as a permanent solution to the problems AI has generated. Nevertheless, approaching the problem in this way would protect authors’ rights to any original contributions they make to AI-assisted works, authors’ interests in the privacy of their creative processes, the interests of authors whose works may have been used without their consent to train AI, and the ability of alleged infringers to challenge the existence and scope of claimed copyrights in AI-generated and AI-assisted works. It would also ease the burden on copyright examiners.
Read full abstract