How are the authorship and ownership of AI art blurred in a resource-sharing platform today?
Nowadays, we are living in a digital age, and digitalization creates a digital culture that is different from all of the cultures that we have before. Indeed, technology has always been improved and digitalization creates new possibilities for our life. Nevertheless, digitalization inevitably creates lots of controversies in reality at the same time. Digital art as a brand new form of art today has caused heated discussions in various fields. How does digitalization create controversies about digital art in sharing culture?
There are two controversies about digital art I would like to discuss, namely authorship blurred in AI art and the ownership lost in sharing culture. The AI creates the possibility to create art by algorithm, which leads to the debate over whether computers can create art and whether AI art has aura and authenticity (Benjamin, 1986/1936; Hertzmann, 2018: Jone, 2016; Zeilinger, 2021). Besides this, we are living in a sharing culture, everyone can easily share or take content, including AI art, from social media platforms, blog posts, or Wikipedia (John, 2017; Fuster Morell, 2011). It challenges the notions of ownership and causes the discussion about amending laws on intellectual property rights for cyberspace.
Certainly, there have already been many positive and negative arguments on AI art and authorship issues. Whereas, I believe that there are two significant elements missing in this heated debate. The first one is the “proportion”. To be more specific, the proportion of computers participating in the process of creating art. It is important to take into account the contribution of non-human units in creating artwork when trying to define if it is real art or not. The second one is cognitive asymmetry. Since there is no clear line to define when the authorship of AI art will be regarded as a computer, everyone has different opinions about the art created by an algorithm. When there is cognitive asymmetry, controversy is expected.
Undoubtedly, digitalization blurs the authorship and ownership of art in sharing culture today. In this paper, I will compare different voices about AI art and take a real case to analyze how cognitive asymmetry play the role in a controversy. Overall, this paper focus on how the authorship and ownership of AI art are blurred in the resource-sharing platform today.
Authorship blurred in AI art
Art can no longer be presented only in physical form, but also in digital form. Further, not only human beings can make art, but computers can also “generate” art these days. The art created by artificial intelligence is called AI art. “Computers creating art” is a brand new technology and a heated issue today.
Some studies have already indicated that computers do not have authorship of art, and AI is not recognized as an author from a legal point of view (Hertzmann, 2018; Zeilinger, 2021). This is because AI art does not have an aura or soul, it is not original nor authentic, and a real painting shows the experience of the author (Benjamin, 1986/1936; Jone, 2016). On the other hand, Zeilinger (2021) pointed out that the algorithm in AI is to make art that fits humans’ standard of aesthetics, and the AI artist Robbie Barra emphasized that AI can be “creative”, and art can be created by an algorithm.
Indeed, the authorship of AI art is vague and difficult to define. The artist could be the user, the algorithm, the programmer of the algorithm, or even the authors of the images in the database. Whereas, the point is “how is the authorship of AI art blurred?” Before answering how the authorship of AI art is blurred, we should look at the development of art in a historical context. Since the invention of pen and paper, people have been using these technologies to make paintings. These technologies are considered tools. Back to the present, people create art from traditional technology to digital technologies. An algorithm is a new tool for expressing and transforming our art and culture (Hertzmann, 2018). No one would judge that the painting made with pens and paper is not “real” art, while plenty of criticisms showed up immediately after the advent of AI art.
It can be seen that the blurring of the authorship of AI art has a great relationship with the proportion of the use of technology. Compare with traditional art, technology is involved too much in the creation process of AI art. Although Hertzmann (2018) pointed out that humans are always the mastermind behind artworks and a computer is simply a tool, when the proportion of tools in the entire creative art is too large, and the participation of human and non-human units is out of balance, the authorship of AI art becomes difficult to define.
Consider the Next Rembrandt, as an example. The Next Rembrandt, which was made by software, was unveiled in 2016. The software system is based on Rembrandt’s paintings to analyze his use of geometry, composition, and painting materials. And then a facial recognition algorithm was used to identify the most typical geometric patterns for human feature painting. With the release of the news, many criticisms came one after another. Jones (2016) criticized that the portrait created by an app has no soul and life, and AI art is a mockery of real art. Additionally, he insisted that the existence of art has historical value, and the paintings can reflect the stories of the time. Since the Next Rembrandt is mainly made by the algorithm and the participation of technology is much more than people are used to, it blurred the authorship of this AI art.
Overall, every time technology participates in the process of making art more than it did, there are voices of doubt, such as filmmaking and photographing were first introduced. The same is true for AI art. In order to understand how the authorship of AI art is blurred, we should focus on the role that technology play in creating art, instead of arguing if a computer can create art. When technology or non-human units are involved in the process of making art to a certain extent, people may start to wonder if it is a real art, thus blurring the authorship of art.
Intellectual property of AI art lost in sharing culture
The technology used in AI art has already made its authorship difficult to define, the rise of sharing culture even makes digital art lose its ownership. On a resource-sharing platform, every user can share and download the content from it. It is already difficult to define the authorship and ownership of general art in digital form, and it is even more troublesome when it involves intellectual property issues of AI art on sharing platforms. This is because there is no clear line between sharing and stealing, and no clear definition of the authorship of AI art. Thus, everyone has different viewpoints and controversy arose in the context of cognitive asymmetry.
Take a controversy that occurred on a platform for creating and sharing images, GANBreeder, which was rebranded as ArtBreeder after this controversy. It is a platform that provides GAN-based image synthesis technology for non-specialist users to create images. GAN, the abbreviation for generative adversarial network, mainly consists of the generator and the discriminator. The generator produces new images, and the discriminator compares them to the images from databases and then decides to reject or validate the output. In the end, the painting that is approved by the discriminator to meet human aesthetic standards will be presented.
On the platform, the users can share their creations and use other users’ creations as source material for further creation. This also had pulled the trigger for the controversy. In 2019, an artist Danielle Baskin accuse another artist Alexander Reben of misappropriating her images without attribution and using them commercially without her permission.
Baskin and Reben both maintained different arguments, and neither side thought they are wrong. Reben believed that the images were generated by an algorithmic system, and were not human-authored. In this case, Reben did not consider those images to be artifacts, and he did not think them would be protected by copyright. On the other hand, Baskin saw herself as the creator of those images, and the computational system was not believed to affect her creative efforts. In her view, the AI system was not perceived as any authorial agency.
In this case, Zeilinger (2021) pointed out that Reben and Baskin have different points of view on the intellectual property of GAN art. When there is no consensus or a clear definition, there is a controversy. The algorithm participating in art blurs the authorship of AI art, and sharing platforms make the ownership of AI art difficult to define. From this example, we can see how ownership of AI art is lost in a sharing platform. There was a disagreement between Baskin and Reben about who had authored the images in contention. To sum up, the ownership of art is not clear in sharing culture, and because of cognitive asymmetry, it may easily cause controversy about intellectual property.
Conclusion
This paper shows how two elements, proportion, and cognitive asymmetry, impact the authorship and ownership of AI art. The interpretation of aesthetics is always changing and it is hard to define. When technologies participate in the process of creating art to a certain extent, when people rely on technologies too much, when tools do too much work for the authors, it makes artworks lose their authorship. Further, when the users do not have a consensus on the copyright of AI art, disputes are unavoidable, and the ownership of the art is easily lost on a sharing platform.
Reference
Hertzmann, A. (2018). Can Computers Create Art? Arts, 7(18), 1–25.
Zeilinger, M. (2021). Generative Adversarial Copy Machines. Culture Machine. Machine Intelligence. 20, 1–23.
Benjamin, W. (1986/1936). The work of art in the age of mechanical reproduction. London, England: Random House.
Jone, J. (2016, Apr 06). The digital Rembrandt: a new way to mock art, made by fools. The Guardian.
https://www.theguardian.com/artanddesign/jonathanjonesblog/2016/apr/06/digital-rembrandt-mock-art-fools
John, N. (2017). Sharing — from ploughshares to file sharing and beyond.
Fuster Morell, M.F. (2011). The Unethics of Sharing: Wikiwashing. International Review of Information Ethics 15, 9–16.
