hankyoreh

Drew Hemment, a professor of data arts & society at the University of Edinburgh.

Drew Hemment is an artist and a professor at the University of Edinburgh’s Futures Institute, an internationally renowned institution in the field of generative artificial intelligence that produced Geoffrey Hinton, considered the godfather of AI.

Using the keywords of art, AI, and data, Hemment has carried out innovative experiments aimed at solving social issues such as the climate crisis, expanding the domain of art in the process. In 2009, Hemment was selected as a fellow of the Royal Society of Arts, a group comprising those who have contributed to the development of human thought.

According to Hemment, artistic meaning has always been fluid, with art changing each time there is an innovation in technology. “We are in a revolutionary moment,” the artist says.

The Hankyoreh conducted an interview with Hemment over email ahead of his participation in the second annual Hankyoreh Human & Digital Forum, held in Seoul on June 16, to talk about changes to how humans practice and experience art and culture in the age of AI.

Drew Hemment: AI gives artists superpowers.

Today we have to contend with deepfakes and misinformation. It can be very hard to disentangle the creations of humans and machines. But we also have phenomenal creative possibilities that would have been unimaginable until just a few years ago.

The advances we’ve seen recently have been made possible by two things — the availability of huge troves of data, and advances in the algorithms and the software that runs these systems. There has been an explosion in powerful new generative AI tools over the past months, such as text-to-image generators like DALL-E 2 and Midjourney, and the ChatGPT chatbot, made possible by recent advances in diffusion models and large language models.

Generative AI is fundamentally changing the ways we interact with machines, and we can expect it to change the way we do things in all sorts of areas, including the arts.

I work with artists who are creating spectacular works with AI. But AI can also generate trivial and uninteresting results. AI tools give humans the capability to find patterns in vast troves of data. But most good art still has a human artist in the loop, someone to ask the profound questions, and introduce human intuition.

Hemment: Artists have experimented with AI, ever since it was created in the 1960s. In recent years, this has exploded, due to advances in technology and the release of powerful new tools.

Today, we see artists working with AI in highly imaginative ways. Artists are at the forefront of pushing the boundaries of the technology.

Artistic meaning has always been fluid, and AI art is no different. It has changed throughout history, with every technological innovation. Without a doubt, we are in a revolutionary moment now.

Today, AI artists are rightfully claiming their place in the history of human achievement.

Hemment: With new capabilities, come big challenges.

AI is what we call “Janus-faced” — it is both very positive and very negative at the same time.

There are a lot of major challenges and controversies with AI. AI can amplify harmful bias in the historic data on which it is trained, and it is energy intensive. Today we have to contend with deepfakes and misinformation, and it can be very hard to disentangle the creations of humans and machines.

One major concern is the current industry-standard AI models have been trained on massive datasets scraped from the Internet, such as images or news. That happens without permission, awareness, or fair pay for the original creators. This still needs to be tested in the courts, but it is obviously plain wrong.

I expect to see a new generation of models trained on licensed images, but that brings a new problem, as only a small number of big players have the scale to do that. That leads to further centralization, with an ever-decreasing group of companies calling the shots.

Hemment: Because of the computing power involved, these systems can spot connections in huge troves of data and so generate insights that the human brain never could.

With AI we will have an “autocomplete” for everything. This will give us all a superhero helper, with unlimited access to data, giving us the ability to generate outputs we could never generate alone. We won’t start with a blank sheet of paper anymore, and we won’t be isolated or alone. Instead, we will be able to build on every image and every word available to our data-hungry friends.

Like any tool, an AI tool can impose a style on artwork created with it. AI art typically looks different, it is what we call “preternatural” and “more than human.” But the outputs of one tool can all look similar, and that is uninteresting to me; it’s not good art.

AI tools combine powerful algorithms and are only as powerful as the data on which they are trained.

Hemment: We have already seen AI-generated artworks sold at Sotheby\'s, and NFTs have turned the art market on its head.

AI is creating huge efficiencies and reducing the need for repetitive tasks in creative workflows, but I am interested more in transformative experiences for audiences, not just in increased efficiency.

I have been working with artists who use AI as a tool, and also as a theme in their work. These are what I call “critical AI artists,” in that they are asking questions of the technology, not taking it at face value, but pushing boundaries, challenging it. I find it is artists working in this way who create the most interesting and inspiring results.

AI is transforming the visual arts, music, gaming, and artists are finding creative new uses for even troubling technologies, like deepfakes.

Hemment: The printing press democratized access to the written word. The personal computer democratized access to data processing. AI tools give more people access to the means to produce artwork. But they are also “black boxed” and make components of the creative process less accessible. So creativity becomes more accessible and democratized, and less accessible and democratized, at the same time.

Hemment: There are three main “creators” in AI: the developer of the technology, the creator of the data on which it is trained, and the artist who uses the tool.

Artists need to be curators and remixers as well as creators of work. Workflows are changing fundamentally. We are being freed from repetitive tasks. And the current tools call for new skills, such as prompt engineering, the art of sculpting prompts, then seeing what the model generates, selecting from a number of options, and reworking the results.

In my team, we build AI tools that give artists more creative control than just prompt engineering. We give artists the ability to directly train and probe a model, without the need to write their own Python code.

Hemment: We need smart and responsible design if we are to get the technologies that can be a positive for humans, and for planet Earth.

I work with artists who combine art and engineering in their practice, so they can build their own tools. A good artist will push a tool to its limits, to extract surprising results.

The New Real provides tools to probe a model so we can better understand how the AI produces what it does, how that is shaped by its training data, and understand the nuance in the difference of the machine intelligence.

Many of the projects I do also have a social purpose, as well as an artistic purpose. So we might be working with climate data to solve a climate problem, at the same time as making new art. The art has a double function: to inspire and to solve problems.

By Han Gui-young, Hankyoreh Human & Digital Institute researcher

Please direct questions or comments to [english@hani.co.kr]