September 26, 2023
In a recent decision, the US District Court for the District of Columbia has ruled that works autonomously generated by AI-based systems shall not be entitled to copyright protection.
Unlike most of the world, copyright creators in the US must register their works in the US Copyright Office in order for copyright protection to apply. Stephen Thaler, the developer and owner of the ‘Creativity Machine’, a GenAI image creation system, has submitted an application to the US Copyright Office for the registration of an image generated by said system, entitled “A Recent Entrance to Paradise”.
The US Copyright Office has denied Thaler’s application on the basis that the work lacked human authorship, which is a prerequisite to support copyright claim. In the abovementioned ruling, the US District Court for the District of Columbia has dismissed Thaler’s appeal with respect to the US decision of Copyright Office.
It should be noted that Thaler tried to challenge the US Copyright Office in order to set a precedence of awarding copyright protection for AI generated works by naming the AI ‘Creativity Machine’ System as the author of the work. However, in view of the court’s demand for human involvement for the provision of copyright protection, Thaler tried to revise his claims during the proceedings of the appeal, claiming that the image was generated by the AI system under his control and instructions, but the court has dismissed such later claims as they directly contradicted the original application submitted by Thaler to the Copyright Office.
Regardless of addressing this specific case, in which the circumstances thereof were not fully presented before the court, this decision sets the foundations of a dual-layered barrier for future of Generative Artificial Intelligence systems by having their generated works lack copyright protection:
Initially, while the decision leaves a wide opening for determining the minimal human involvement for copyright protection entitlement, we should keep in mind that such human involvement in providing prompts is traditionally not such that entitles copyright protection. Copyright is aimed to protect manners of expression, both US and Israeli Copyright Acts clearly state that ideas are not entitled to copyright protection, only the form of their expression. When prompts are provided to AI-based systems such as ChatGPT and Midjourney, these prompts merely instruct such systems as for the idea the works should reflect and express, while the manner of their reflection and expression are mostly controlled by the systems themselves, thus not allowing any human involvement in the generation process that would be sufficient to entitle copyright protection.
Secondly, the Thaler decision completely shuts down the option of providing copyright protection to works autonomously generated by AI-based systems, which are likely to be developed in the near future in many categories, such as music and video AI-based platforms. In case music streaming platforms will add features for music generation based on the users’ playlist and general taste, the generated are will not be copyright protected.
Legislators should jump in and face this issue, in a manner that would both prevent setting impossible barriers for this technology, and would recognize and benefit the humans who created the data which AI-based systems need to create new works autonomously.
This document is intended to provide only a general background regarding this matter. This document should not be regarded as setting out binding legal advice but rather as a practical overview that is based on our understanding. APM & Co. is not licensed to practice law outside of Israel.