The emergence of accessible generative AI products has resurfaced the question of the degree of human creativity that is required for copyright protection. While the Copyright Office has a study underway, both the recent guidance from the Copyright Office and a recent case holding suggest that human creativity continues to be an integral requirement in securing copyright protection. Though the Copyright Office conducts a case-by-case analysis with a thorough review of the facts, without a human providing some modicum of creativity in the traditional elements of authorship, a work is not likely to be granted protection for the foreseeable future.
I. Background
A. Copyright Authorship Requires Originality
According to the Copyright Office, copyright is a “type of intellectual property that protects original works of authorship as soon as an author fixes the work in a tangible form of expression.” Notably, the protection of a copyright applies to a broad range of works, from traditional books to works of art such as photographs, poems, paintings, musical compositions, and the like.
Originality is actually a constitutional requirement because the power to enact copyright laws stems from the Copyright Clause (Article I, § 8, cl. 8). To qualify as an original work, the work must be created by a human author and include some degree of creativity. In other words, in order to secure copyright protection, the work at issue cannot simply be copied. There is no statutory definition of original works of authorship, but it is well-settled by the Supreme Court that the degree of creativity required for copyright protection includes some modicum of creativity. The Trademark Cases, 100 U.S. 82 (1879). The question of human authorship has been addressed with respect to celestial beings, monkeys, and gardens. In each case, courts declined to extend copyright protection because human authorship was lacking.
B. The Recent Popularization of Generative AI
Generative AI has become recently popularized due to the introduction of highly accessible tools such as ChatGPT and Midjourney. However, generative AI has a long history with roots tracing as far back as the 1950’s. While generative AI was in full motion back in the 90s, the computer power required was a barrier to providing mainstream solutions.
The continued strides in technological developments paved the way for the modern generative AI solutions that have flooded the market in the past couple of years. As a result, there are amply more works created using AI now than ever before. Users supply prompts, whether simple or complex, to the AI tool and receive an incredibly complex output, often differentiated from the baseline training data used to train the AI. Through these AI tools, users are generating writings, images, and the like. Thus, while generative AI is not new, the influx of created works leveraging AI through newly accessible mainstream tools is resurfacing the question of the degree of human creativity required for copyright protection.
II. Recent Developments
Human authorship was recently addressed in Thaler v. Perlmutter. At issue was whether the claimant’s copyright application was properly denied because the visual art in question was generated by AI. Thaler v. Perlmutter, No. CV 22-1564 (BAH), 2023 WL 5333236, at *3 (D.D.C. Aug. 18, 2023). In holding that the application was properly denied, the court reiterated the long-standing notion that human authorship is a bedrock requirement for copyright protection. Id. Similarly, the Copyright Office canceled the copyright registration on a graphic novel that was developed by entering prompts into Midjourney because human authorship was lacking.
Equally important, in the first quarter of 2023, the U.S. Copyright Office provided guidance in response to the popularization of newly accessible AI tools. The Copyright Office reiterated authorship requires human input and that merely mechanical or machine-produced works devoid of human creativity are un-copyrightable material. Applicants must disclose the use of AI and provide a brief explanation of the humans’ involvement. The Copyright Office guidance indicates AI-generated content may be copyright protected if the work is in some way original and unique. However, if the traditional elements of authorship were generated by a machine, then the work lacks human authorship, and the Copyright Office will not register it. The Copyright Office conducts a case-by-case, specific inquiry, in which examiners are looking for works that demonstrate an original mental conception by a human being.
Lawmakers have called for the creation of a national commission to address whether changes are needed to existing law to address the evolving landscape of AI. With this Congressional focus on the implications of AI and innovation, it is likely an area of policy that will continue to develop. While the U.S. Patent and Trademark Office is requesting comments regarding AI ownership, the U.S. Copyright Office has launched an initiative to examine policies affecting AI. The Copyright Office has issued a Notice of Inquiry in the Federal Register, undertaking a study of policy issues raised by generative AI and existing copyright law. Furthermore, the Copyright Office has recently extended the comment period, allowing initial written comments from the public through October 30, 2023.
III. Conclusion
It’s important to remember that any proposed changes to existing copyright law will still be subject to the constitutional principle of originality under the Copyright Clause as a constitutional amendment is unlikely. Even with prospective changes on the horizon, originality will likely remain a threshold requirement in seeking copyright protection. Copyright protection is unlikely to be extended to a work generated by simple AI prompts. Accordingly, keeping humans in the driver’s seat on the traditional elements of authorship is likely the best course of action for copyright protection for the foreseeable future.
About the Author: Sarah Covington is an Associate in the Firm’s Intellectual Property practice group. She holds an MBA and previously served as Vice President of Revenue Cycle and as COO for a healthcare technology start-up focused on creating AI solutions. Ms. Covington has a particular interest in computer software coding, having led three teams to win at MIT hack-a-thons with AI-based health solutions. She has received awards for emerging AI solutions at the Health Data Palooza and National Policy Conference. In 2022, Ms. Covington received the Rising Star Award from the South Dakota Bar Association.