U.S. Copyright Office Ponders Copyright Protection for AI-Generated Works Following D.C. Circuit’s Thaler Decision
The U.S. Copyright Office (“Office”) has issued a notice of inquiry (NOI) in the Federal Register indicating its intent to study copyright law and policy issues raised by generative artificial intelligence (AI), and seeking public comment from interested stakeholders. The Notice follows on the heels of the D.C. Circuit’s recent decision in Thaler vs. Perlmutter, upholding the Office’s denial of an application for protection of visual AI-generated art, on the grounds that the art “lacked the human authorship necessary to support a copyright claim.”
According to the Notice, the Office has received numerous applications to register works containing AI-generated material, and naming AI systems as an author or co-author in recent years. At the same time, there has been an increase in copyright infringement claims brought against AI companies for their use of protected materials to train AI models. The Notice is one part of a larger AI Initiative announced by the Office earlier this year, which included a Statement of Policy on AI issued by the Office this spring regarding its position on the copyrightability of certain AI-derived outputs.
Public comments in response to the Office’s Notice are due by Oct. 18.
From the Typewriter to Thaler
Since at least 1965, the Office has been active in issues regarding machine learning and copyright protection. According to the Notice, the Office has always declined to take a categorical position on registration of machine-derived work, with its former Register noting at the time: ‘‘a typewriter is a machine that is used in the creation of a manuscript[,] but this does not result in the manuscript being uncopyrightable.’’ The copyright inquiry, dependent as it is on “circumstances of a work’s creation,” requires a nuanced evaluation of authorship; only those original works of authorship are capable of protection under the Copyright Act. To date, the Office and federal courts have limited copyright protection to works of human authorship: “non-human spiritual beings,” living gardens, and monkeys have all had copyright applications submitted on their behalf rejected.
But what of works that are derived from some mix of human and AI involvement? To put it another way, what level of human involvement is required to obtain copyright protection over a work? Two recent actions by the Office provide insight into where the Office may draw the line:
- Zarya of the Dawn: Zarya of the Dawn is a graphic novel with mixed text and visuals. The visual graphics, including the novel’s front page and interior imagery, were generated by the publicly available AI art generator Midjourney. After first registering the work, the Office discovered (through the applicant’s Twitter feed) that the images were generated by Midjourney. The Office issued a notice to show cause why the registration should not be canceled given the applicant’s failure to disclose the use of Midjourney AI, and the applicant responded. Ultimately, the Office revised the registration allowing copyright protection for certain portions of the work, specifically the novel’s “text,” as well as the novel’s “selection, coordination and arrangement” of both text and images. The images themselves, however, were deemed not capable of copyright, as they were generated by Midjourney AI and were not the product of human authorship.
- Thaler’s autonomous Creativity Machine application: Stephen Thaler is an AI advocate on a crusade to expand patent and copyright protection to AI-derived content. In a patent case, Thaler has argued that his “Device for the Autonomous Bootstrapping of Unified Sentience” should be considered an “inventor” for patent purposes (a position that was rejected by the US Patent and Trade Office and IP regulators in other countries), and elsewhere Thaler has argued for recognition of AI as a sentient being. Thaler also owns a computer system he calls the “Creativity Machine,” which can generate original pieces of visual art. Thaler tried to register one of these works called, “A Recent Entrance to Paradise” and listed only the Creativity Machine as the author. Thaler argued that the copyright should then be transferred to him, as a work-for-hire owner of the machine. The Office denied Thaler’s application in full, on the grounds that the visual art “lacked the human authorship necessary to support a copyright claim.”
The Office’s actions on Zarya and Creativity Machine confirm the Office’s policy that autonomously generated AI imagery – whether through proprietary AI tools, Midjourney, ChatGPT, or DALLE – are not capable of copyright protection. Further, the Office indicated a willingness to parse an application to apportion protection to human-derived content, separate from its AI counterparts.
Thaler’s District Court Challenge
Thaler challenged the Office’s decision twice, claiming that the AI should be acknowledged “as an author where it otherwise meets authorship criteria, with any copyright ownership vesting in the AI’s owner.” After the Office denied both of those challenges, refusing to register a claim where a human being did not create the work, Thaler challenged the Office’s decision in the U.S. District Court for the District of Columbia. Among other things, Thaler challenged the Office’s denial under the Administrative Procedures Act, arguing the denial was “arbitrary, capricious, an abuse of discretion and not in accordance with the law, unsupported by substantial evidence, and in excess of Defendants’ statutory authority.”
In its analysis, the District Court found that the single legal question presented was whether a work generated autonomously by AI is capable of copyright protection. Again, there was no dispute that the Creativity Machine alone, and not in conjunction with Thaler, generated the artistic image at issue.
The District Court affirmed the Office’s denial, holding that: “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, even as human creativity is channeled through new tools or into new media. Copyright has never stretched so far, however, as to protect works generated by new forms of technology operating absent any guiding human hand.” The Court did not address Thaler’s other theories on ownership, as the work at issue did not give rise to a valid copyright capable of transfer in the first place.
The Thaler Court was aware of the precedential impact of its decision on the future of AI and copyright law, noting:
“Undoubtedly, we are approaching new frontiers in copyright as artists put AI in their toolbox to be used in the generation of new visual and other artistic works. The increased attenuation of human creativity from the actual generation of the final work will prompt challenging questions regarding how much human input is necessary to qualify the user of an AI system as an ‘author’ of a generated work, the scope of the protection obtained over the resultant image, how to assess the originality of AI-generated works where the systems may have been trained on unknown pre-existing works, how copyright might best be used to incentivize creative works involving AI and more.”
What’s Next for AI and Copyright?
As articulated in the Statement, the Office will continue to evaluate applications with AI-generated content with the following principles in mind:
- Works made “without any creative contribution from a human actor” are not copyrightable;
- A work including “human-authored elements combined with AI-generated images” is copyrightable, although the individual AI-generated images are not;
- When an AI “receives solely a prompt from a human and produces complex written, visual, or musical works in response, … users do not exercise ultimate creative control” and the resulting work is not copyrightable;
- When a human selects, arranges, or modifies AI-generated material in a sufficiently creative way, the work may be copyrightable;
- A user can use technology to transform or adapt their otherwise expressive authorship without necessarily negating a human-driven creative process.
Together, the Office’s Statement, decisions in Zarya and Creativity Machine, and its ongoing AI initiative efforts will continue to work to define what the Office views as the outer limits of copyright protection in the AI/ML space. But applicants like Thaler and others will continue to test these limits in new and creative ways as commercial use of AI-generated content rapidly increases across industries. Applicants should be mindful of this guidance and seek competent counsel before submitting applications that include AI-generated outputs.
Clark Hill’s Artificial Intelligence and Emerging Technologies team advises clients on emerging challenges related to the use of AI/ML tools. For further information, please reach out to the authors or a member of the AI/ET team.
This publication is intended for general informational purposes only and does not constitute legal advice or a solicitation to provide legal services. The information in this publication is not intended to create, and receipt of it does not constitute, a lawyer-client relationship. Readers should not act upon this information without seeking professional legal counsel. The views and opinions expressed herein represent those of the individual author only and are not necessarily the views of Clark Hill PLC. Although we attempt to ensure that postings on our website are complete, accurate, and up to date, we assume no responsibility for their completeness, accuracy, or timeliness.
WEBINAR: The Race to 2024: Politics and Social Media in the Workplace and Employer Rights.
Over the last several years, employers have seen and continue to see increased political activities from their employees at work and on social media platforms, including on business-related social media platforms, like LinkedIn. Managing employee expression causes unique challenges for employers and HR professionals, and in a General Election year, these challenges are likely to increase as the Presidential race, and other races, heat up.
Webinar: A Cookieless Future and Promise of PETs: A Primer on Privacy Enhancing Technologies
This webinar will explore PETs – we will define what they are, what problems PETs exist to address, and emerging PET standards including the National Institute of Standards and Technology (NIST) draft guidance on how to evaluate PET effectiveness. We will provide specific PET use cases and discuss how PETs may be utilized to address the phase out of third party cookies by certain browsers for purposes of targeted advertising.
WEBINAR: Cybersecurity Resilience in Law Firms
This webinar focuses on law firms seeking useful information about robust cybersecurity strategies to protect their clients, maintain ethical and legal compliance, and fortify their digital infrastructure.