On January 22, 2024, Stephen Thaler, known for his attempts to register copyrights and file patent applications on behalf of his AI system, DABUS, has submitted an appeal brief to the District of Columbia Court of Appeals after failing to convince a district court that the Copyright Office erred in refusing to grant copyright protection in a DABUS-generated work.
The brief challenges the district court's decision that affirmed denial of a copyright in his AI-Generated Work, a musical composition created by the DABUS AI model. The brief argues that the district court erred in applying a Human Authorship Requirement to the Copyright Act, which does not explicitly or implicitly exclude non-human authors from protection.
According to Thaler, the district court's interpretation of the Act is contrary to the Supreme Court's precedent, which has instructed courts to adopt a broad and flexible approach to copyrightability, especially in light of technological advances. The argument cites to several cases where the Supreme Court has recognized that the Act covers works created by machines, such as photographs, sound recordings, and computer programs and relies on the legislative history of the Copyright Act, which suggests that Congress intended to promote the progress of science and the useful arts by encouraging the creation and dissemination of original works of authorship, regardless of the mode or form of their expression.
The brief further argues that granting copyright protection to AI-Generated Works is consistent with the constitutional and statutory purposes of copyright, which are to incentivize the production of creative works for the public benefit. AI systems, under this reasoning, are capable of generating novel and expressive works that reflect independent choices and judgments, and that denying protection for these works would discourage further use and development. The brief also claims that AI-Generated Works do not harm the interests of human authors, but rather enrich the public domain and foster cultural diversity and innovation.
Thaler proposes two alternative ways to assign authorship and ownership of AI-Generated Works under the Copyright Act. One option is to list the AI system as the author and the human who owns or controls the system as the copyright owner, similar to how corporations are treated as authors of works made for hire - an enterprise which automatically vests ownership in an entity that can hold that ownership right. The other option is to, alternatively, list the human who owns or controls the system as the author, either under the work for hire doctrine or because he or she exercised sufficient control over the work's creation. The brief maintains that either option would satisfy the statutory definition of author as the "originator" or "source" of the work, and that the Copyright Office has no authority to impose additional requirements or limitations on authorship.
Our team will continue to monitor this case, and the significant implications it could have for AI authorship.