The first topic of an open discussion as part of the AI Hokusai Art & Tech Research Project was “The Ethics of AI in the Arts.” Our Open Discussions are an opportunity to explore and debate how AI is reshaping the art world and participate in a global dialogue at the intersection of technology and creativity.
Speakers of the discussion were:
• DAVID GUNKEL - researcher, author, professor of technology ethics and aesthetics;
• PAUL ROSERO CONTRERAS - researcher, interdisciplinary artist, professor of Art, Science & Technology;
• SAINT DENIS - new media artist, merging AI and VR in immersive art and installations;
• SARES - digital artist blending traditional aesthetics with 3D, AI, and generative art;
• ALEX MAY - multimedia artist known for his innovative exploration of the intersection between art, science, and technology;
• ANNA SHVETS - moderator and curator of AI Hokusai ArtTech Research project.
We are publishing some parts of the webinar to encourage you to watch it in full on our YouTube channel.
The discussion started with the story of the relationship between artificial intelligence and arts, as presented by David Gunkel:
So, no matter what devices you're talking about—typewriters, books, film projectors, cameras, electric guitars, you name it—these are all considered tools. Our technologies are considered tools in the hands of human creators: human artists, writers, painters, filmmakers, musicians, etc. This idea of technology as tools is very old; it dates back to Plato's Phaedrus in the ancient world. However, I think it gets its best direct modern expression from Martin Heidegger, a German philosopher. Heidegger calls this the instrumental theory of technology. We could say that instrumental theory has been instrumental in helping us understand technology and its role not only in the world of art but also in all kinds of human activity. We understand technology as a tool, but I think we are at a moment in time now, in the late 20th and early 21st centuries, where we are confronting a new normal: the machine is moving from the position of a tool to that of an active or interactive agent in the creative process. This development presents us with new opportunities and some new challenges that I believe are important for us to grapple with.
The large language model is a perfect way of making that theoretical idea practical. We now have texts that are generated without an author. These texts are unauthorized. The question then becomes: how do we arrange for questions of responsibility to be answered in the face of text generators like OpenAI's ChatGPT, Google's BERT, or Anthropic's Claude? We have the same issue about image generators, whether it be Mid Journey, Crayon, DALL-E 2, or Stability AI. We can now train large models on all the digital images that human beings have ever created. The algorithm can create its own 'original' images. The questions arise about who is the artist and what it means when we use previous artworks to create new content that could compete with human artists. A great deal is currently happening in the legal realm as people try to resolve copyright and intellectual property questions regarding training these models on copyrighted content available on the Internet. We also see developments in the audio sector; Udio, Suno, and Speechify are doing the same with audio data because all media is now digital.
All media serves as training data, as Holly Herndon and Matt Dryhurst have recently stated. When all media becomes training data, it allows us to create powerful new technologies for content generation. However, it also raises many questions about responsibility, accountability, and ownership over the work of art, including who is accountable for it or able to control it. This issue is not recent; we've been observing its evolution for quite some time. In 2018, the Obvious Collective in Paris created a portrait of Edmond Bellamy in a series of portraits they generated using a GAN, a generative artificial neural network. They put it on sale at Christie's, and it sold for half a million dollars. This marked the entry of AI-generated art into the art auction world.
We are in a peculiar moment where we see the tool wanting to be perceived as a co-author, a co-collaborator, or a co-creator. This opens a new way of thinking about technology and its role in art practice and our process. However, this issue extends beyond art and artistry; it challenges our feelings. The problem is not simply whether computers, machine learning, algorithms, or other applications can or cannot be responsible for their действия; it also relates to how we have determined, described, and defined artistic responsibility in the first place.
Then we continued the discussion with Paul Rosero. Anna Shvets asked Paul, "As an interdisciplinary artist, how do you balance your scientific and creative aspects when integrating AI?"
I have some projects where we're using AI in different ways, in topics related to archaeology or marine ecology. In archaeology, when you don't have actual evidence, data, or something that physically supports your proposition or the hypothesis you are looking for, we found a space—some room—to introduce the use of artificial intelligence, to try to, as it was mentioned before, use it as a tool, something that helps us fill this gap, historically speaking, where we don't have certainty or evidence to propose something. In this example, we were working together with Fernando Studio, an equatorial archaeologist, on a project in the Galapagos Islands, researching the first people inhabiting the islands. There aren't many photographic archives or much information related to those years. In this case, we had the chance to introduce generative artificial intelligence to produce animations to tell the story of how this first settlement on the island occurred.
A new media artist, Saint Denis, also joined the discussion.
Let us know if you have any moral issues with the artworks created with AI, - asked Anna.
The topic that I want to cover in terms of ethics is censorship and privacy in AI. The digital era will change not only the tools we have but also the way we think in general. One of the fundamental changes of the new digital era is cloud technology. Everything we are creating right now is online; it's blockchain. Two years ago, I mainly worked locally on my computer, and now everything I do is online. We have Google Colab, network, and notebook technologies—a lot of stuff. Most of the tools with AI require an Internet connection, and the work is now being done somewhere outside of my machine; it's being done on some servers.
When you're working online, it's an entirely different question of privacy you have, and why privacy is crucial for artists. Over 100 years ago, Malevich worked on his copy concept of Suprematism art. Before his first exhibition, where he presented this idea for two years, he hid his works in his studio because it was crucial for him to be the first one. He needed to create a surprise effect; imagine how this work changes when you're working online. Why was privacy so crucial for Malevich? If you delete privacy, here comes censorship. The main difference between working locally on your computer and online is that you must be prepared to share your work at any stage.
Anna also asked a digital artist, SARES, about his experience with AI:
Have you ever been blamed for working with AI? Has somebody told you you are not a real artist because you do everything on your computer with artificial intelligence? Do you just have some buttons, and AI does all the rest?
I’m sure that someone has once said that. There’s a significant point in my case because I came from 3D and 3D technologies, and 3D artists sometimes don’t like the tool. I understand that because the production pipeline changed a lot. However, I saw it as an opportunity because, for example, you can push and explore the process more without overthinking about the craft. Like, I don’t know, spending three months or one month modeling a plate, a cup, or a glass, whatever. You can explore more within the piece. Regarding the art part of AI, I think there are two big differences. We can sell AI as a content producer, which is a really, really good tool. I use it a lot in my other projects. But here we are talking about art, and that’s where I think the limitations and censorship are not good in my opinion. Our job as artists is to express ourselves, and we should not try to limit that expression regarding censorship or what I can do or should do versus what I shouldn’t do.
Later, Alex May, our other resident and a very talented artist who works at the intersection of art, science, and technology, shared his vision on ethics in art created with AI:
I sat down with ChatGPT and spent two hours feeding it all of the questions I had. I then asked it to generate a 10-minute talk, which I’m going to read to you. Today, I want to delve into a subject that’s been increasingly occupying my thoughts: the ethics of artificial intelligence in the arts. It’s a timely and deeply personal topic to many of us in the creative community. I wrote this script with the assistance of ChatGPT, which seems quite fitting given our discussion today. As artists, we’re navigating an era where the intersection of technology and creativity is more prominent than ever before. AI tools offer incredible possibilities, but they also present ethical dilemmas that we need to address thoughtfully. First, let's consider the dependency on AI tools. Artists have always relied on tools from the paintbrush to the camera, but the tools we’re discussing now are fundamentally different.
Art has the power to be transformational and transcendent on an individual level. Focusing solely on maximizing audience appeal, especially through AI-generated art designed to cater to mainstream tastes, can undermine this potential. The private, personal engagement with art is something that can’t be mass-produced. It’s a dialogue between the artwork and the individual—a space for introspection and emotional resonance. We must ensure that in embracing AI, we don’t lose sight of the profound impact that art can have on a personal level. In conclusion, as we stand at the intersection of art and artificial intelligence, we face both incredible opportunities and significant ethical challenges. It’s up to us as artists, technologists, and audience members to navigate this landscape thoughtfully. We need to advocate for the responsible use of AI that respects the artist's craft, protects intellectual property, and preserves the transformative power of art.