AI Text-to-Image Art: Can We Actually Protect it with the Indonesian Copyright Law?
The development of GAN (Generative Adversarial Network) model, DALL-E, Contrastive Language-Image Pre-training (CLIP), and Diffusion Model, plays a huge role in developing the text-to-image art generators. OpenAI is one of many companies that are currently developing these futuristic features, including their newest DALL-E 3 text-to-image generator that can translate nuanced requests with prompt texts into extremely detailed and accurate images. There are a lot of different types of AI generators (e.g. ChatGPT as a text generator) but now let’s focus on the AI text-to-image generator because a lot of people don’t really know the existence of these features, but it’s affecting a lot of creative workers (illustrators, graphic designers, etc).
As aforementioned, text-to-image art generator has a lot of different models and it’s developing as time goes by. One of the most developed and commonly used by OpenAI, MidJourney, and Imagen is the Diffusion Model. In understanding the IP protection aspect of the AI-generated image arts, let’s dive into how the Diffusion Model works generally.
After the user inputs the prompt text, the first step is the training data, which is simply a dataset of collecting all images based on the alt texts of these images scattered throughout the internet
When all word descriptions have been collected in the dataset, the next stage is through deep learning. AI tries to classify the prompt texts according to the visual concept that best matches the prompt text, this process is simply like teaching AI to mix and match millions of inputs in the form of algorithm code and this process is carried out by AI in the form of deep learning.
The next stage is latent space. All the variables that have been collected in the deep learning process then begin to be formed in the latent space.
Keep in mind that the variables collected by deep learning obtained from the dataset are in the form of pixels, so in the latent space, there is a mathematical space that can determine whether the text description inputted by the user has a round/square shape, green/orange color, or has a shadow/not, and other indications that indicate 1D/2D/3D or even more than three dimensions (multi-dimensional latent space). The illustration of the latent space above is the process where each pixel-shaped variable is measured for its corresponding position.
After the AI forms pixels with the formula according to the prompt text instructed by the user in the latent space, the pixel in the form of a mathematical matrix is converted into an image, this process is called diffusion.
It’s all very technical and seems to be ‘confusing’ for people who are not familiar with AI processing. However, what we can conclude is that the user (as the human) is not involved directly in the process of the AI-generated image art. Users input the prompt text and wait for AI to process it into an image or an illustration as the user requested. The question is: regardless of the prompt text by the user, can we say that the user is the creator/originator of the art?
The question regarding the authorship of the artwork determines the possibility for such artwork to be protected by copyright law or not. In Regulation Number 28 of 2014 on Copyright (“Indonesian Copyright Law”), the answer is very simple and straightforward: the image art created by AI cannot be protected by the Indonesian Copyright Law, because in order for the works to be protectable, it has to be created by the “natural” author.
The definition of Author according to Art. 1 of the Indonesian Copyright Law means:
a person or several persons who individually or jointly produce works that are unique and personal.
Thus, the mindset of the Indonesian Copyright Law (which was established in 2014) didn’t take into account the capability of AI to create creative works that are protectable under the copyright law.
Furthermore, the originality of the works supports the concept of authorship in the copyright law. The Berne Convention does not provide an explicit definition of originality. However, Indonesia applies the civil law doctrine, whereas according to Goldstein and Hugenholtz, in contrast to the common law (such work ‘originated’ with the author and not someone else), civil law countries usually view originality more specifically based on the imprint of the creator’s personality on the work. The Indonesian Copyright Law does not further define originality in the nomenclature, however, the words implied from “unique and personal” as stated in the definition of Author emphasized originality in such works.
To elaborate further on the originality of the works, we have to look into the U.S. doctrine of copyright law. In the case of Feist v Rural (1991), the judge requires two things in proving originality in a work: it requires independent effort and a modicum of creativity. The independent effort can be interpreted as an independent effort to generate intellectual works that do not copy other authors’ creations, notwithstanding originality is not equal to novelty, as long as such works have the “independent effort and personality” that distinguish such works from others. The modicum of creativity emphasizes creative choice as part of the creative process, whereas creative choice has a sense of freedom in making the works according to our personal touch and our independent effort
Back to the context of AI text-to-image art, independent effort and a modicum of creativity determine a work’s originality. The AI text-to-image art is generated through the process of deep learning by AI (dependent on AI’s interpretation), with the ‘creative choice’ also from the AI’s interpretation of the dataset collected from the internet. The user doesn’t have much of a creative choice nor independence to determine the direction of the artwork, regardless of the user’s effort to input the prompt text. Despite the popularity of the ‘art’ of prompt text in AI-text-to-image art where users get to explore different kinds of techniques of prompt text to create better illustrations/images, it is debatable that prompt text is part of the user’s creative choice as the creative effort between an illustrator and a user with a prompt text is different.
If the originality of the AI-text-to-image art cannot be fulfilled, then the work also does not fulfill the element of authorship according to the Indonesian Copyright Law doctrine. Therefore, the AI cannot be considered as the Author, nor can the work itself be protected by the Indonesian Copyright Law, which causes the copyright ownership in such work is not recognizable. In the context of ownership under the Indonesian Civil Law, the user may possess such works as an intangible asset but does not necessarily have the ownership of such creation or the “eigendom” right (Art. 570 of Indonesian Civil Code). This means the user cannot enjoy the possession with economic rights and moral rights given by the copyright law as the owner of the creation, hence the ownership of the AI-generated image art can never be claimed under the Indonesian Copyright Law.
However, if the AI is not recognized as the Author, can the human who created the AI be recognized as the creator of the work?
The civil law doctrine does not recognize such concept, so does The U.S. (as a common law country). We can see this from the Stephen Thaler case in 2018. Thaler applied copyright for “A Recent Entrance to Paradise”, a work by Creative Machine, an AI-driven art generator machine. Thaler wanted to register this computer-generated work as work-for-hire to the owner of the Creative Machine. However, the U.S. Copyright Office rejected the registration on the grounds that it lacked the human authorship necessary to support a copyright claim and work-for-hire doctrine is based on the employment agreement between employer and employee (natural person), hence such doctrine cannot be implemented in the case.
The U.S. is different from most common law countries, as the UK, Ireland, Hong Kong, India, New Zealand, and South Africa protect computer or AI-generated works as creations that can be protected as intellectual property. As stipulated in the “U.K. Copyright, Designs, and Patent Act 1988”, “South Africa Copyright Act”, and “New Zealand Copyright Act”:
in the case of a literary, dramatic, musical, or artistic work that is computer-generated, the person by whom the arrangements necessary for the creation of the work are undertaken.
With the comparison of computer-generated works protection between common law doctrine and civil law doctrine above, we can conclude that there’s a difference in the perspective of authorship and originality between common law doctrine and civil law doctrine. This difference is caused by the history of copyright itself, which will be another subject to uncover in another article.
In conclusion, with emphasizing the originality of the creation, the AI-text-to-image art is not eligible as a protected work under the Indonesian Copyright Law. However, with the development of technology and rapid awareness of artificial intelligence in this era, we don’t know if this doctrine will always exist as the basis for addressing this phenomenon. There will come a time when AI as a creator/originator is a common phenomenon, and a doctrine is needed for humans as the creator of such AI to be recognized as the creator of such AI-generated works — and we, as Indonesians, have to be ready for such changes.
