No Human Authorship, No Copyright

Intellectual Property, Technology & Software

by | Sep 19, 2023

“Copyright is designed to adapt with the times. Underlying that adaptability, however, has been a consistent understanding that human creativity is the sine qua non at the core of copyrightability”, said the District Court for the District of Columbia in Stephen Thaler v. Shira Perlmutter, et al. 2023 U.S. Dist. LEXIS 145823 (D.C. Cir. 2023) (“Thaler”), an opinion that came down recently on August 18, 2023. The Thaler court found that copyright protection does not extend to works generated by Artificial Intelligence (“AI”) – i.e., absent proof of human authorship, a work is simply not copyrightable. However, with the rise of AI, what constitutes “human authorship” may not be so clear.

The Thaler Decision

In Thaler, the plaintiff used a computer system – which he referred to as the “Creativity Machine” – to generate a work of visual art that he later sought to register for a copyright. The computer system was listed as the author in the plaintiff’s application to register the computer-generated work. The plaintiff claimed, however, that the copyright should nonetheless transfer to him as the owner of the Creativity Machine. The U.S. Copyright Office denied the application, finding that because the computer-generated work was not authored by a human, it could not be protected by copyright law. The plaintiff contested the denial, suing the U.S. Copyright Office in the District Court for the District of Columbia. The parties in Thaler ultimately brought dueling motions for summary judgment. In ruling on those motions, the court was tasked with deciding whether a work generated entirely by an AI system, absent any human involvement, should be eligible for copyright. According to Thaler, the answer to that inquiry is no – for now. The Thaler court’s holding that copyright protection does not apply to AI-generated works is grounded in the express language of the Constitution, The Copyright Act of 1909 (the “1909 Act”), The Copyright Act of 1976 (the “1976 Act”), the congressional intent underlying those acts, and Supreme Court authority.

Article 1, Section 8, Clause 8 of the Constitution allows for the enactment of copyright law by granting Congress the power “to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” From its inception, it was thought that copyright law could and should be used as a tool to incentivize human beings to create and invent in a way that would benefit the public good, by granting human authors exclusive rights in their original works. As highlighted by the Thaler court, “non-human actors” simply do not need to be incentivized in that way because they are machines, and thus, copyright law could not have been designed to apply to AI systems.

Consistent with the express language of the Constitution, the 1909 Act states that “any person entitled thereto… may secure copyright for his work” – keyword, “person”. Under the 1909 Act, it was clear that to obtain a copyright for a work, the work must have been authored by a natural person. And, importantly, according to the Thaler court, there is no indication that Congress intended to alter or remove that express human authorship requirement from the 1976 Act. The Thaler court pointed to a congressional report (H.R. REP. NO. 94-1476 (1976)) to support that finding, which states “Congress intended to incorporate” into the 1976 Act “the ‘original work of authorship’ standard ‘without change’ from the previous 1909 Act.” The human authorship requirement, then, was explicitly built into both the 1909 Act and the 1976 Act.

The Supreme Court has also consistently acknowledged and enforced the human authorship requirement. In Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884), Mazer v. Stein, 347 U.S. 201 (1954), and Goldstein v. California, 412 U.S. 546 (1973), the Supreme Court made clear that there is a direct correlation between the level of human creativity and the copyrightability of a work. The more human involvement that goes into a work, the more likely it is to be copyrightable. For example, the Supreme Court’s analysis of Sarony is particularly informative. In Sarony, the Supreme Court found that photographs are copyrightable creations despite the fact they are technically generated by a mechanical device because photographs “represent the original intellectual conceptions of the author.” In other words, photographs – although captured by a device – could not be brought to life without some degree of human creativity. Sarony is evidence that the human authorship requirement can still be met where a human uses technology to assist with the creation of a work, so long as that technology is used as a mere conduit to bring the human’s original creative vision to life. Following in the footsteps of Sarony, the Thaler court distinguished between a human utilizing a mechanical device to produce a work that still reflects the human’s creative vision (which is copyrightable), and a human pressing a button or two on an AI system to generate a work that is not the product of the human’s creativity (which is not copyrightable).

The Future of the Human Authorship Requirement

The facts at issue in Thaler were cut and dry. The work at issue was generated entirely by an AI computer system, which the plaintiff did not dispute. The plaintiff listed the “Creativity Machine” as the author of the work in his registration application. That begs the question – what about the copyrightability of a work that is generated by a human, with the assistance of AI? And, similarly, what about the copyrightability of a work that is produced in part by a human author, and in part by AI? The Thaler court was not faced with addressing those questions, but it did recognize that those questions are going to arise in future cases, and to answer them, courts may need to consider redefining what constitutes human authorship.

The Thaler court acknowledged the “new frontiers” we are approaching when it comes to AI and its impact on intellectual property law. According to Thaler, that new frontier will require courts going forward to determine:

  1. How much human involvement is required to qualify a user of an AI system as an author under the 1976 Act, such that the work is copyrightable?
  2. The scope of protection afforded to a work that is generated with the assistance of AI.
  3. How to assess the originality of works generated by AI.
  4. How copyright law can be used to incentivize creative works that utilize AI, while still serving the public policy underlying copyright law – i.e., incentivizing the creation of original works in a way that benefits the public good.

It is not a matter of if – but when – courts will be called upon to decide those questions. While the human authorship requirement has historically been the bedrock of copyright law, the rise in the use of AI to generate artistic, visual, and musical works cannot be ignored. As technology continues to develop, it is only going to become more commonplace for people to utilize AI systems to create and invent. That is significant because if AI-generated material is not copyrightable, people may be deterred from seeking copyright registration for a large majority of works. What that means for copyright law and the human authorship requirement remains to be seen – perhaps there will come a time when determining whether the human authorship requirement is met will require a more nuanced analysis than simply distinguishing between human-authored works and AI-generated works.

Guidance from the U.S. Copyright Office

On March 16, 2023—several months before the court issued its decision in Thaler – the U.S. Copyright Office issued a Statement of Policy, in which the U.S. Copyright Office sought to clarify its policies related to examining and registering copyright works that contain AI-generated material. In the Statement of Policy, the U.S. Copyright Office explains that it begins by asking the following: is the work one of human authorship, with AI merely being used as an assisting instrument, or were the traditional elements of authorship conceived and executed entirely by AI?

To illustrate, the U.S. Copyright Office makes the following distinction – if a human submits a prompt to an AI system asking it to create a work, but the AI system has full control over the expressive elements of its output, such work is simply not the product of human authorship and is not copyrightable. Conversely, if a human author arranges AI-generated material in a particular way or otherwise modifies material originally generated by AI in a sufficiently creative way, the human authorship requirement may be met, at least to some extent. For example, the U.S. Copyright Office recently found that a graphic novel consisting of both human-authored text and AI-generated images was a copyrightable work, however, the images themselves were not subject to copyright because they were the product of AI and not human authorship. The human-authored text, however, was copyrightable.

The Statement of Policy gives applicants some insight into how the U.S. Copyright Office is currently assessing the copyrightability of AI-generated works, and what to expect when trying to register a work for copyright that contains AI-generated material. The Statement of Policy offers applicants the following tips:

  1. Individuals applying for copyright protection must disclose to the S. Copyright Office whether any portion of a work submitted for registration was created by AI and specify the human author’s contributions to the work.
  2. If applicants seek registration for a work consisting of both human-authored and AI-generated elements, applicants should exclude the portions of the work generated by AI from the registration application (applicants may do so by listing the AI-generated portions of the work in the “Limitation of the Claim” section of the registration application).
  3. Applicants should not list AI as being the author or co-author of a work they are seeking to register for copyright.

In light of the Statement of Policy, it seems as though the U.S. Copyright Office and at least some courts agree that AI-generated material is simply not copyrightable, because as it stands, the human authorship requirement prevents copyright law from extending to such lengths.

Takeaways

Works generated entirely by AI, with no human creativity or involvement, are not copyrightable. The Statement of Policy issued by the U.S. Copyright Office earlier this year and the court’s decision in Thaler make that clear. However, as the line between human-authored works and AI-generated works becomes less clear, the intellectual property laws that have been well-solidified for decades also become less clear. Courts will undoubtedly have to further evaluate the human authorship requirement as AI becomes more prevalent, but for now, copyright protection simply does not extend to AI-generated works. Individuals seeking copyright registration for a work generated with the assistance of AI should expect that copyright protection will only extend to the portions of the work that are proven to be human-authored. If you want to ensure that your entire work is copyrightable, you may want to think twice about using AI to assist with the creation of that work. If AI is used in any capacity, you should be careful to disclose that fact in your application for copyright registration and specify to what extent AI was used.

*Not intended to constitute legal advice to be relied upon. TALG recommends that you consult with legal counsel to receive tailored advice suitable to your circumstances.

Author

  • Cassidy Kitterman

    Cassidy Kitterman joined TALG in 2022 as an Associate. Cassidy was born and raised in Portland, Oregon. She later moved to Southern California to attend the University of San Diego, where she received her undergraduate degree in Theology and Religious Studies, cum laude.