All points
1745 words
- Refik Anadol is a Turkish-American pioneer of AI art. The 40-year-old artist has spent more than 16 years combining data, artificial intelligence and visual creativity. He describes his approach as painting with a “thinking brush,” using information as an artistic medium rather than relying on conventional drawing or painting techniques.
- Anadol's Los Angeles studio demonstrates how immersive AI art works. The space features enormous LED screens, mirrors and constantly evolving images generated by artificial intelligence. The visuals transform shapes, colours and textures, creating an immersive environment that resembles a constantly changing digital landscape.
- The experience extends beyond visual art into other senses. Visitors wear a device around their necks that releases AI-generated scents, including rain and flowers, to accompany the images. Anadol plans to introduce another device that monitors visitors' vital signs, including heart rates, and uses that information to change the artwork in real time.
- Anadol sees immersive AI installations as a new artistic medium, not a technological party trick. He describes the experience as discovering a new place that has never existed before. His ambition is to create environments that go beyond what conventional artistic methods can produce.
- His work has appeared on some of the world's most prominent architectural and entertainment structures. Major installations have been displayed on the Sphere in Las Vegas, the Walt Disney Concert Hall in Los Angeles and Antoni Gaudí's Casa Batlló in Barcelona.
- Anadol considers data a form of artistic pigment. Unlike conventional paint, which must dry, data can continually change its shape, colour, texture and form. He believes information surrounding humanity has its own voice and can become the raw material for artistic expression.
- One of his projects used 200 million photographs of Earth. NASA data provided the foundation for an artwork that transformed enormous quantities of planetary information into visual experiences. The project demonstrates how scientific datasets can become material for creative work.
- His California landscape project involved 153 million curated images. Anadol demonstrated the system using a gaming controller, allowing him to navigate and manipulate the digital environment. The collection included images of natural subjects such as flora and fauna.
- His AI production process converts photographs into mathematical representations. Each image becomes a collection of data points describing characteristics such as colour, texture and shape. These representations are plotted in multidimensional space, where an AI system learns their relationships and characteristics.
- The AI generates new images from patterns learned from its training data. When given a prompt, the system produces new visual combinations that do not represent actual physical locations. Anadol describes these imaginary environments as worlds that exist in the machine's mind and can subsequently be reshaped through human interaction.
- Anadol applies additional algorithms to establish his distinctive visual style. These algorithms blend and transform the generated images into continuously flowing compositions. His artistic contribution therefore extends beyond selecting prompts to designing and controlling the processes that determine how the images evolve.
- Anadol deliberately aims for a 50/50 relationship between human and machine creativity. He says he has spent approximately ten years trying to achieve a balance in which human decisions and machine-generated results contribute equally. He sees himself as an artist who uses computers rather than a programmer who happens to make images.
- His approach has attracted major technology companies and institutional partnerships. Anadol has worked with Google, MIT and Microsoft on large public installations. His collaborations have helped establish AI-generated visual art within prominent technological and cultural institutions.
- Anadol's work has achieved substantial commercial recognition. His pieces have sold for upwards of US$1 million at auction, and his installations have been exhibited in museums internationally. This places his work within the high-value contemporary art market despite continuing disagreements about whether AI-generated images qualify as art.
- The Museum of Modern Art in New York commissioned Anadol's installation Unsupervised in 2022. The approximately 24-foot-high installation occupied MoMA's lobby and reimagined the museum's collection through artificial intelligence. Its scale and prominent placement represented significant institutional recognition of AI-based artistic experimentation.
- Former MoMA director Glenn Lowry explained how Unsupervised was created. Anadol trained an AI system using publicly available metadata from MoMA's entire collection. The metadata described the characteristics and identities of artworks, providing the digital information through which the system could reinterpret approximately 200 years of art.
- The algorithms behind Unsupervised transformed information from existing artworks into entirely new visual combinations. Data associated with one object could evolve into the data of another, producing a third or fourth object that had never previously existed. Lowry considered the resulting experience deeply satisfying.
- Lowry described Unsupervised as an extraordinary public success. Visitors reportedly sat in front of the installation for hours, transfixed by the changing images. He said it became one of the most popular images in MoMA's collection.
- Anadol claimed visitors spent an average of 38 minutes with Unsupervised. This compares with approximately 28 seconds that studies reportedly found museum visitors typically spend looking at major artworks. The difference became one of the strongest arguments offered for the installation's ability to hold public attention.
- Pulitzer Prize-winning New York Magazine art critic Jerry Saltz strongly disputed the idea that popularity establishes artistic success. He described Unsupervised as a giant lava lamp or an expensive screensaver, comparing its constantly changing imagery to an Etch A Sketch. He reportedly valued the installation at the level of a half-million-dollar screensaver rather than a significant artistic achievement.
- Saltz believes meaningful art should provoke personal reflection rather than merely hold attention. He argued that spending 38 minutes looking at an installation does not demonstrate artistic achievement. Great art should make viewers become quiet within themselves, confront uncomfortable questions and explore unfamiliar emotional or intellectual territory.
- Saltz does not reject AI art as a category, but believes its development is immature. He said AI art will eventually become art, while describing much of the current output as poor quality. His estimate was that approximately 90% of AI-generated art was rubbish, but he added that approximately 90% of Renaissance-era art was also rubbish. His argument was that new artistic mediums require time to mature and that challenging work is often initially misunderstood.
- Lowry compared the controversy surrounding AI art with the historical arrival of photography. Photography challenged traditional assumptions about artistic creation because the human hand was no longer directly responsible for producing every image. He sees a similar challenge emerging with artificial intelligence and believes technological progress cannot simply be stopped.
- Artist and author Molly Crabapple considers AI-generated art trained on other artists' work a major violation of creative ownership. She described the practice as the greatest art heist in history, arguing that AI companies had taken billions of images rather than stealing individual paintings from museums.
- Crabapple believes museums, galleries and auction houses should refuse to support AI art trained on artists' work without their permission. She argues that institutions should not purchase or exhibit work produced through systems that commercially exploit other people's creative output without consent.
- Crabapple describes popular AI image generators as corporate plagiarism bots. She argues that these systems are trained using artwork scraped from the internet, including her own illustrations, without artists being consulted or compensated.
- A demonstration involving Crabapple's illustration of Aleppo, Syria, illustrated her concerns. The programme asked an AI image generator to create a Syrian skyline in the style of Molly Crabapple. The system produced a visually similar illustration within seconds, raising questions about how closely AI-generated work can reproduce an identifiable artist's style.
- Crabapple says artists have generally received neither consent requests nor compensation or credit for the use of their work in AI training datasets. When asked whether anyone had requested permission to feed her images into such systems, she said no. She also argued that artists often cannot even see where their work has been used.
- AI companies defend their use of copyrighted material under the doctrine of fair use. They have told lawmakers that their systems study and learn from copyrighted works in a way comparable to human learning. Fair use allows copyrighted material to be used without permission in certain circumstances, although its applicability to AI training remains disputed.
- A group of artists has filed a class-action lawsuit against four AI image-generator companies. The artists accuse the companies of copyright infringement and other wrongdoing connected with the use of their creative work. The dispute brings the conflict between AI development and artists' intellectual property rights into the legal system.
- Anadol acknowledged the concerns raised by artists about unauthorised use of creative work. He said he understood why his artist friends described the practice as theft and agreed with their underlying concerns. However, he distinguished his own production methods from the practices being criticised.
- Anadol says he has used only ethically sourced datasets since 2020. He described permission as the most important part of creating AI art. His team researches datasets, establishes where the information originated and begins its work by obtaining permission to use the material.
- Anadol is building a 20,000-square-foot museum dedicated to AI art in downtown Los Angeles. The institution is called Dataland and is intended to provide a substantial physical space for celebrating AI-generated artistic experiences. The project represents a major commitment to establishing AI art as a recognised cultural discipline.
- Dataland reflects Anadol's belief that AI should expand artistic possibilities rather than threaten human creativity. He wants to create experiences that no individual human could produce alone, using technology to extend the boundaries of imagination.
- Anadol acknowledged that artificial intelligence lacks human emotion, lived experience and genuine intent. When these limitations were raised, he agreed that they were real characteristics of current AI systems. Rather than denying them, he argued that human-machine collaboration is the way forward.
- Anadol's ultimate position is that humans and machines should develop art together under responsible conditions. He believes this collaboration can bridge the gap between human imagination and machine capabilities, provided that the resulting work is shared fairly and created responsibly.
- The central disagreement remains unresolved: whether AI-generated work is genuinely art or an appropriation of human creativity. Anadol sees a new artistic medium built on collaboration and responsibly sourced data; Lowry sees a continuation of earlier technological transformations; Saltz questions the artistic depth of much current AI work; and Crabapple challenges the consent and ownership practices underpinning commercial AI generators.