As part of the fourth DigiCon annual conference, themed “Ownership and Digital Constitutionalism,” the Sci-Fi section organised a special workshop, Generative Art and the Future of Ownership: Perspectives from Creators. This workshop explored the perspectives of DigiCon’s artist community—including Ivo Witteveen, Gordon Johnstone, Andrei Stan, Fillip and Michał Bedkowski—on GenAI, focusing on critical areas such as property rights, their artistic approach to GenAI as well as Generative Art. We have divided the workshop content into two posts. In this first post, our sci-fi editor, Etienne Valk, introduces Tools vs Totems, which delves into the topics discussed in the first half of the video. The aim of this workshop is to understand how creators are responding to these shifts as generative AI impacts the creative industries, while also providing lawyers with insights directly from those affected.
“How can she make a song written by a machine sound so beautiful?” the protagonist wonders aloud in the film adaption of George Orwell’s 1984 from 1949. The idea of machines producing art is nothing new, nor is putting that idea into practice. The scale on which this is made possible and by which it is used is something of recent years.
A 2024 survey found that a staggering 92% of content creators worldwide are already using Generative AI (GenAI) tools in their creation process. While this figure should be taken with a grain of salt, especially when compared to other survey outputs from the same year, it’s interesting that 72% of creators expressed concerns about the ethical and appropriate use of the technology.
When exploring how GenAI is perceived, the most evident group to look at first are arguably the creators themselves. Within the endlessly diverse group of creators that make up the world, responses regarding GenAI range from extreme fear to extreme excitement and often a mix of both, as the above figures seem to suggest.
As part of The DigiConference IV, we held a workshop with creators from different parts of the western world who are active in varying fields of the visual arts and music sector. We integrated their perspectives and input on questions regarding generative art and the future of ownership in this piece. For a video summary of the workshop, see above.
Totem or tool?
Sometimes, GenAI is being personified in the general public. Glorifying qualities are being attributed to it. Were you to use a search term online like “the next [insert archetypal artist] might just be a robot”, you shouldn’t be surprised there are a few hits of news articles. But it seems that the closer people are to either the building or the application of the big data-driven models, the clearer it becomes that if GenAI had any persona at all, it would be that of a will-less servant. One that is especially useful to employ for automating and thus accelerating processes, on command. So, while GenAI might present an appearance of agency or intention, it doesn’t actually possess any.
Here, it’s important to briefly distinguish between the many possible uses GenAI can have. Probably the most eye-catching is the generating of complete pieces of music, visual art or written work, usually based on a human prompt. Other common examples are the enhancement of existing work into new, sometimes endless work, generating background images in video editing, finetuning the writing of prose or the mix of a piece of music, and the splitting of a song into separate parts.
In the survey mentioned in the introduction, one-third of the creators using GenAI stated saving time was one of the three main reasons for their usage of the technology. Looking at the examples above, in most cases, it’s clear that time is being saved. The efficiency is maybe even most obvious when generating entire pieces of work, while only entering a few words as a prompt. Seconds as opposed to hours, days or months.
Still, it could make one wonder about the meaning of creation – putting feelings, experiences or ideas into form.
Especially considering the following.
Another third of the creators in the survey listed generating new ideas among their top three reasons for using GenAI. If the idea for a work is generated, then materialized, and perhaps finetuned, all through GenAI, you could still argue that it’s only a tool in the hand of a human creator. But you could also argue the tool is put on top of a totem pole, with a human standing underneath, both hands put together, hoping to catch something fruitful falling down.
In other words, aspects of the technology might have an inherent, perverse incentive to prefer laziness over effort for (some of) its users. Gradually relying more on the glorified technology and less on themselves.
Legal implications also come to the fore here. If there’s too little new, human input to a work, copyright protection isn’t given. At least under EU and US jurisdiction, a human author and originality are central requirements of a copyrightable work.
Value of the output
Ownership and the value of creative output are expressed more broadly in the creative sector than ‘just’ through copyright. Complex questions arise about the relationship between the value of the creation process and the value of the output of that process. Yet another third of the surveyed creators say one reason they use GenAI is to produce higher-quality content. On the more extreme part of the spectrum of attitudes, Chinese artist Ai Weiwei said that art that can be easily reproduced by GenAI is “meaningless” and that that kind of art should have died a long time ago. There are creators with more nuanced takes. One of the workshop participants, Dutch music composer and producer Ivo Witteveen, mentioned during the DigiCon workshop: “creating music is not about saving time, it is about spending time in a fulfilling way.”
This is also telling coming from a creator making a living from music on a commission basis, which generally means being paid per delivered piece of work. For primarily commercial works, as opposed to primarily artistic works, it might be easier to see that GenAI output is more efficient, time and money-wise, and thus preferable. Still, there are several points made within the creative sector that push back against the idea of promoting automation when it comes to mainly supportive, “functional” creative content.
One point is that creators can sustain their more creative input by getting income from their commercial activities. If companies were to turn to GenAI directly, important revenue streams would pass by creators.
Another practical point is that when employing GenAI tools (in their current state), it’s much more uncertain for an employer whether the GenAI delivers copyright-infringing output. Although that possibility also exists with a human creator, the human creator would generally be liable, while liability would lie with the employer in case it uses GenAI.
An important point brought forward is that creators can sharpen their craft and try out ideas in their commercial pursuits, which can then enrich their artistic endeavors. But also, why settle for creative output, wherever it lies on the spectrum of artistic and commercial, that’s merely ‘acceptable’ at best? Music composer and producer Witteveen expands on this as follows: “When an AI tool helps me with creating a mix, I can judge whether it is doing a good job. But I can only do that, because I have been mixing music for a long time. I have been trained to listen to a level of detail and nuance that most listeners haven’t. But what if I would have started out my career with using AI mix assistants? How would I be able to discern if it ‘sounds OK’ or ‘sounds perfect!’?”.
The last point particularly ties in with the feared “devaluation” of art as a consequence of GenAI. Here, devaluation isn’t just meant as art bringing in less money because of a GenAI flood on online platforms that submerges existing work, which gives rise to enough concern in itself.
It’s important to also consider the devaluation because of a reduction in the quality of the art.
Whether the current quality of primarily generated output is on par with primarily human-made output according to ‘the public’ is not entirely clear yet. We do know a risk is that the more omnipresent GenAI works become on the internet, the more new GenAI output will train on and be “contaminated” by those works. As a 2023 study puts it: “The Curse of Recursion: Training on Generated Data Makes Models Forget”.
Despite rapid developments within AI, the question remains whether GenAI, without meaningful human interference, could ever really come close to the human art of creation. As Australian singer-songwriter Nick Cave orated in his open letter about GenAI: “This is what we humble humans can offer, that AI can only mimic, the transcendent journey of the artist that forever grapples with his or her own shortcomings. This is where human genius resides, deeply embedded within, yet reaching beyond, those limitations.”
An interesting concept, however, is the idea of the generative models themselves as pieces of art. Elaborate models can come about through an endless number of design and data input choices, which in turn will decide how the output will look or sound. Visual artist and machine learning engineer Fillip Isgro expects that the molding of their own intricate, personalized models will become an important part of the creative process for many artists in the (near) future. Of course, such wider adoption would require a more accessible and malleable backend of GenAI tools, and possibly more data and programming proficiency – plus willingness – on the side of creators.
Outlook
As Isgro concurs, on what data, or rather, on which creative works those models are trained, is of great importance. Just last week, more than 30,000 artists and other people from the creative industry signed a petition declaring against the wrongful use of creative works. Their statement read as follows: “The unlicensed use of creative works for training generative AI is a major, unjust threat to the livelihoods of the people behind those works, and must not be permitted.” According to a 2024 US survey, a majority of creators express concerns about AI’s impact on the art workforce and who profits from their art. At the same time, around 45% of the creators surveyed agree that “AI art models are a positive development in the field of art”. Creative people are often at the forefront of society and not afraid of changes to the status quo. However, they also tend to criticize where criticism’s due.
Within the GenAI debate, Mexican filmmaker Guillermo del Toro sharply noted: “The value of art is not how much it costs and how little effort it requires, it’s how much would you risk to be in its presence?” For many companies the answer is: a whole lot of lawsuits.
Etienne Valk
Has a background in music, sociology, and law.
Editor at DigiCon (sci-fi section)
