Generative artificial intelligence tools produce a vast range of new content, including code, text, audio, images and video. For the business user, the speed of output in response to a user prompt can deliver game-changing business efficiencies. However, the appeal of generative AI content needs to be balanced against the implications of using that content within a business. There are several dimensions to this, and one important question – with interesting potential outcomes – is the extent to which the user can own the output. Superficially, this might seem to depend on the terms of service of the particular generative AI platform and the allocation of rights set out in its governing terms. While this is indeed a part of the story, the user also needs to consider whether that output is even capable of being owned, by anyone, under applicable law. Below, we explore the answer to that question around the world, based on law and guidance as of the date of this post.
US: No ownership
The default position under US law is that copyrights in creative works of authorship fixed in a tangible medium vest in the author immediately upon creation (17 US Code §201). However, US copyright law has repeatedly been interpreted to require human authorship for that ownership, such as in Naruto v. Slater, also known as the “monkey selfie” case, where the US Court of Appeals for the Ninth Circuit held in 2018 that a monkey does not own the copyright in a photograph it snapped of itself. In 2023, the US District Court for the District of Columbia reaffirmed in Thaler v. Perlmutter the absence of ownership for works generated by an artificial intelligence tool that plaintiff Stephen Thaler developed, although in this fact pattern, Thaler intentionally limited any human creativity and emphasized the role of the machine. Notably, the series of case law and guidance from the US Copyright Office, which is currently our primary source of guidance for individuals seeking copyright ownership of works they developed with the assistance of generative AI tools, establishes that, under current US law, there is no ownership of AI generated works by anyone – not by the authors of the AI tool, not by the tool itself and not by the individual who enters the prompts to generate the work. As such, these works are currently considered to be in the public domain, without copyright protection.
The US Copyright Office issued further guidance explaining that “a work containing AI-generated material will also contain sufficient human authorship to support a copyright claim,” where, for example, an author had made creative arrangement of AI-generated works, or substantially modified AI-generated works. However, “[i]n these cases, copyright will only protect the human-authored aspects of the work, which are ‘independent of’ and do ‘not affect’ the copyright status of the AI-generated material itself.” Thus, the US Copyright Office requires identification of AI-generated content and human-generated content in works submitted for registration, such that protection will be granted to the human-authored content only. In its registration decision regarding the comic book “Zarya of the Dawn,” for example, the US Copyright Office denied protection for images created using the generative AI art platform Midjourney, but it allowed registration for the text, and the selection and arrangement of images and text, where the applicant, Kris Kashtanova, attested to sole responsibility for those elements. In the same guidance statement, the US Copyright Office also reaffirmed that protection remains for underlying original works, even where they’re enhanced by technological tools.
The US Copyright Office noted in its guidance that it will continue to monitor this new technology and may issue further guidance in the future. As such, the current guidance around “sufficient human authorship” ultimately could support a future award of copyright protection for AI-generated works, or portions or variations thereof. For now, however, there is no assurance of protection for AI generated works in the US.
European Union: Ownership possible
EU copyright law is a patchwork of 13 directives and two regulations. However, none of this legislation, nor the upcoming EU AI Act, directly addresses the ownership of AI-generated works, and, outside the legislation, there is little in terms of relevant EU-level case law. The Court of Justice of the European Union (CJEU) does provide some limited directional guidance in Infopaq International A/S v Danske Dagblades Forening (Case C-5/08), where it held that copyright will only subsist if there is originality flowing from the “author’s own intellectual creation.” This has been widely interpreted to mean that a significant form of human input is required. Nevertheless, it will be for individual EU member states to determine whether the output of an AI-generative model can meet this requirement. By way of example of the state of play, the German Copyright Act requires an author’s “own intellectual creation” for the existence of a copyrightable work – and it is thought that neither a machine nor a computer program can be the author, so it is presupposed that an “intellectual creation” must be created by a human. Likewise in France, the current presumption is that only natural persons can be considered authors, and originality requires “the personal touch or intellectual effort” of the author, whereas “implementation of automatic and constraining logic” without “genuine personal effort” will not qualify.
However, similar to the above discussion, it remains to be seen where there could be sufficient human input to generate a work with the assistance of an AI-generative model, such that the human could claim ownership of the work.
UK: Ownership available
The UK’s position is similar to the EU’s position, requiring a copyright work to be “the author’s own intellectual creation” and exhibiting an author’s “personal touch.” As in other jurisdictions, copyright will exist where a human author uses a tool, such as a word processing package or a pen, to produce a work. However, if all the “creativity” takes place within an AI platform, it might be concluded that – as with other countries – output from generative AI would not be protected in the UK.
Significantly, however, the UK under its copyright legislation – Copyright Designs and Patents Act 1988 (CDPA) – also extends copyright protection to “computer-generated works.” Although theorists have challenged the idea that a non-human “computer” can generate a copyrightable work embodying creative skill, Section 9(3) of the CDPA clearly provides that the person who made the “arrangements necessary for the creation of the work” is the author of the resulting copyrightable work. This position was recently ratified by the UK government, as well as the UK Intellectual Property Office – which in 2022 held an open consultation specifically on the application of Section 9(3) to generative AI, with the conclusion that Section 9(3) should remain.
So, while there is an open question as to who the person making “necessary arrangements” is in practice, it seems clear that, in the UK, copyright subsists in generative AI outputs.
China: Ownership possible
The general rule in China is similar to the other jurisdictions that we have looked at in this post: A “work” eligible for copyright protection under the Copyright Law of the People’s Republic of China (CL) must be an original “intellectual achievement.” Although the CL does not specifically address the copyrightability of AI-generated content, two recent court decisions are instructive, and might provide a conceptual framework applicable in other countries.
In 2019, the Shenzhen Nanshan District Court found that generative AI output – or some output, at least – may be eligible for copyright protection. In this case (Shenzhen Tencent v. Shanghai Yingxun), the court confirmed that AI-generated text generated using Tencent’s Dreamwriter writing AI software can constitute protectable works. The court found that the “work” in question was not generated “purely by AI,” and that “intellectual activity” existed in the output because Tencent’s inputs, in areas such as arrangement and selection of data input and trigger condition setting, were key to the output generation process. Although Dreamwriter did not have “personhood,” it was relatively straightforward for the court to find that multiple teams within Tencent were behind the work, and that Tencent therefore owned the copyright.
In a more recent case, the Beijing Internet Court ruled that an image generated by Stable Diffusion met the “intellectual achievement” and “originality” criteria because the user made an intellectual contribution by inputting prompt texts and setting parameters, with the result that the AI-generated image reflected a personalized expression of the user. In the same case, the court held that the copyright owner of the AI-generated image was the user, rather than the platform developer, because the developer did not have the intent to create the image and also did not determine the inputs.
These decisions do leave open the question of whether a work generated “purely by AI” could be found copyrightable, but the general reasoning relating to “arrangement and selection” of inputs may provide a path toward copyright protection for at least some users of generative AI tools.
Conclusion
In sum, copyright ownership of AI generated works varies jurisdictionally, where laws in some cases are still being settled. If ownership and protectability is important to the user of a generative AI tool, it is therefore critical to consider the terms offered by the provider of the tool, the governing law of the agreement, the jurisdiction where the user seeks to enforce its copyrights, and other jurisdictional factors.
