Share This Article
A recent denial of protection by the US Copyright Office of images generated by an artificial intelligence system was followed by more detailed guidance recognizing some exceptions.
The denial of protection by the US Copyright Office of works generated by artificial intelligence
A recent decision by the US Copyright Office has denied registration of AI-generated images, sparking a debate about the extent of human creativity and intervention required for legal protection under US Copyright law.
The case centers around Kristina Kashtanova, author of the graphic novel “Zarya of the Dawn,” who used the artificial intelligence image generation software Midjourney to create the images in her book. However, the US Copyright Office ruled that since the images were not created by a human being but rather by a generative AI tool, they are unprotectable under copyright law.
While Ms. Kashtanova was granted registration of her graphic novel in 2022, the Copyright Office later asked for clarification on the extent of human intervention in the creation of the images. The Office reiterated that registration is only possible if an original work was created by a human being, protecting only the result of intellectual labor that relies “on the creative powers of the mind.” On the contrary, works made by a machine or mechanical process that operates randomly or automatically without sufficient creative input or intervention by a human author cannot be granted registration.
Although Ms. Kashtanova was recognized as the author of some elements of her work (e.g. the text drafted with no AI support), the individual images generated by the artificial intelligence were not considered the product of her human creativity and were therefore not protectable under copyright law. Indeed, the Midjourney software independently generates the results obtained, with the user’s prompts resulting in four different images based on Midjourney’s training data. As such, the process by which a user obtains a final image cannot be considered the same as the creative process of a human artist, writer, or photographer.
Consequently, the US Copyright Office canceled the original certificate of registration for “Zarya of the Dawn” and issued a new one covering only the material created by Kashtanova.
The US Copyright Office Guidance on the registration of works containing material generated by AI
The above-mentioned decision was followed by a guidance from the US Copyright Office setting out limits in which works containing material generated by artificial intelligence can be registered.
According to the guidance, when determining the extent to which a human is the author of a work containing AI-generated material, the US Copyright Office asks whether the traditional elements of authorship in the work were actually conceived and executed by a machine or a human. If the work’s traditional elements of authorship were produced by a machine, the work lacks human authorship and is not registerable.
However, if a work containing artificial intelligence-generated material also includes sufficient human authorship to support a copyright claim, the 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. The policy does not mean that technological tools cannot be part of the creative process, as authors have long used such tools to create their works or to recast, transform, or adapt their expressive authorship. Ultimately, the extent of human creative control over a work’s expression determines whether it is registerable under copyright law.
The Italian precedent and the lessons learnt from the position of the US Copyright Office on AI generated works
It is worth it to mention that, unlike what happens in Europe, the US copyright system requires the registration of works to be granted some types of copyright protection. As such, copyright is tested in Europe only in case of disputes.
In this respect, a recent decision of the Italian Supreme Court may be of interest. The case pertained to the copyright protection of the scenography of the Sanremo Music Festival which had been created through a software program that would have processed its shape, details and colors through mathematical algorithms, reproducing and repeating its shapes on different scales of magnitude. The author’s process would, therefore, have been limited to choosing an algorithm and approving the result generated by the software.
In this case, the Italian Supreme Court pointed out that the use of software does not exclude the processing of an intellectual work protectable by copyright, the rate of creativity of which would only have to be scrutinized more rigorously. In any case, if such an exception had been noted before the courts of merit, a more rigorous factual assessment would have been necessary to verify whether and to what extent the tool absorbed the creative elaboration of the artist who used it.
The above mentioned cases subject of the scrutiny by the US Copyright Office and the Italian Supreme Court are substantially different and the AI technology used by Ms. Kashtanova was more advanced that the algorithms used by for the Sanremo scenography. However, these two decisions are an indicator of how difficult is likely to become for authorities to identify copyright protection in AI generated works.
In this context, it will be pivotal for authors to map out and record all the activities performed to create the works with the support of artificial intelligence to be able to defend against challenges.
On a similar topic, the following article may be of interest “Unlocking the Potential of Generative Artificial Intelligence (AI): Navigating the Legal Issues and Unleashing Its Creativity”.
Authors: Giulio Coraggio, Roberto Valenti, Lara Mastrangelo and Rebecca Rossi of DLA Piper