09
DTC 101
Digital Art & Expression
Digital art refers to art that incorporates digital technologies in the making, presentation, circulation, or experience of the work. Because all media can be converted into digital data, digital art includes remediations of older forms such as painting, photography, cinema, music, and poetry, as well as newer hybrids such as creative coding, net art, virtual reality, augmented reality, interactive installations, generative systems, and AI art.
Digital art is not defined only by the tools used to make it. It is also a way of thinking about art as process, system, interface, data, network, and interaction. A digital artwork may be a finished image, but it may also be a program that runs, a website that changes, a simulation that responds, a performance with code, a dataset transformed into sound, or an AI system guided by human prompts and curation.
The inclusive term Digital Art and Expression recognizes that creativity now happens across many spaces: museums, galleries, browsers, mobile phones, social media feeds, game engines, archives, and AI platforms. Digital artists use contemporary technologies not only to make new kinds of images and experiences, but also to question how technologies shape perception, memory, identity, labor, ownership, and culture.
“The computer is not a tool, although it can act like many tools. It is the first metamedium.”
— Alan Kay and Adele Goldberg, “Personal Dynamic Media” (1977)
9.1
Computers and Art
“The Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform…. Its province is to assist us in making available what we are already acquainted with.” - Ada Lovelace
In 1842, Ada Lovelace, inspired by Babbage's research into his Analytical Engine, began taking notes about the creative and not purely mathematical possibilities of computation. She understood that a computer could do what one asked of it. For example, with a computer one could "...compose elaborate and scientific pieces of music of any degree of complexity or extent." Ada's intuitive understanding of code as a tool for complex semiotic processing - a tool for literary, visual and musical creation - contributed to the idea of the computer as a personal creative assistant.
Computer Art began in the 1950s as the creative play of computer engineers and scientists, because they were the only ones who had access to the computers. From these first experiments, visual artists began exploring the possibilities of computers and animation. Algorithms, as Ada had observed over a century earlier, could process any data - incuding dots, lines and shapes. John Whitney is considered to be one of the fathers of computer animation. In the early 60's, Whitney used a self-built analog computer to create animated title sequences and commercials for movies and television. A record of these visual effects he collected in a work called Catalog.
By the 1970s, Whitney began working with faster, digital computers to make work such as his psychedelic digital film Arabesque. These "primitive" works, in comparison to today's computer animation, show the remarkable ability of a computer, with human instructions, to plot change over time, to simulate organic movements and to organize elements into complex patterns. Like cinema, computers have the breath of life.
Language artists also found inspiration in computer programming. bpNichol (Barrie Phillip Nichol) was a Canadian poet, fiction writer, digital and sound poet, editor and publisher. In 1983-1884, he created a series of computer programmed kinetic poems using an Apple IIe computer and the Apple BASIC programming language. When Basic became obsolete, a student replicated the poems in Hypercard. When Hypercard became obsolete the work was rescued with a video screen capture. See FIRST SCREENING below.
Other early computer artists also helped establish the computer as a creative medium. Vera Molnár used algorithms and plotters to explore geometric variation, while Harold Cohen developed AARON, a drawing program that generated original images for decades. These artists are important because they show that questions now associated with AI art—authorship, automation, variation, and collaboration with a machine—were already present in earlier computer art.
9.2
Creative Coding
Artists who hand-code their art are directly working with the unique aspects of a computer's environment: random access, simulation, looping, stored data, user input, complex algorithms, duplication. Not all digital artists are coders, however. A digital photographer works with computer hardware and software and rarely needs to look at code to manipulate an image. Those artists and amateur creators who rely on common digital tools for creating images and sounds, are still creatively manipulating code. A digital "brush" that lays paint on a "canvas" is simply a precise set of instructions in the software, activated by user inputs and passed to the computer for execution. Instead of coding directly with a programming language, a majority of digital creators work with graphic user interfaces to perform specific tasks that the software then translates into code.
The following categories of creative coding do not follow any official taxonomy. These categories are just some of the often overlapping digital and computational techniques that take advantage of binary code and the computer's affordances.
Algorithmic Design
Algorithmic art refers to works that involve a process based on an algorithm, or detailed recipe, devised by the artist for the design and execution of the work. Weaving and knitting are algorithmic arts, because they are processes that involve the repetition of steps. In the 1960's, Fluxus and Conceptual artists gave up the visual object and instead wrote instructions to "execute" the work using the viewer's own physical and mental processes. While these artists meant to provoke and challenge notions of art, they were also responding to the algorithmic possibilities of machines and humans. Algorithms are deterministic. Unless there is some novel input, such as a random number or external data, an algorithm has exactly the same result each time its executes. Avant-garde composer and Fluxus artist John Cage composed many of his works with simple instructions for the musicians and often included some external variable, such as the weather or a throw of dice. His most famous work, 4'33", instructs the performer to sit quietly at the piano for 4 minutes and 33 seconds, allowing the natural sounds of the environment to be the music for a listening audience.
Algorithmic art with computers can produce fantastically complex and unpredicable patterns out of the repetition of a few simple steps written by the artist-coder, along with non-deterministic actions or external variables. Rafael Rozendaal is an algorithmic artist who works with HTML, CSS and JavaScript. He is also one of the first artists to sell websites as art objects. In 2013, Rozendaal’s www.ifnoyes.com website was sold to a collector for $3,500. This work reveals a set of geometric shapes that change size, color and position through the subtle interactions and movements of part to whole. This behavior is confirmed by the user interacting with the mouse: altering individual shapes alters all the other shapes by chains of cause and effect. The deterministc algorithm that causes lines to move, is disturbed by the variability of mouse movements.
Simulation
In general, a simulation is an imitation of a process or system. Because computers are very good at imitating processes, the term "computer simulation" has come to encompass almost any computer-based representation. A computer simulates language processing, because it can only process binary code. In fact, any reference to the external world must be "virtual" - just a description of the world as semiotic output. A personal computer has an interface that models a desktop with paper-looking documents, stacked files, folders and trash can. The process this virtual desktop "simulates" is "working a desk." Virtual simulations have become very useful in studying complex forces within natural and human systems as well as in creating the illusion of cinematic reality in games and movies. Scientists, game developers and artists create complex computer simulations by first modeling 3D environments and then simulating the interactions between elements in those environments. A computer animation of wind blowing through tree branches, would first require the 3D modeling of trees, along with individual branches and leaves. An algorithm for a directional force affecting these branches and leaves would then be written to simulate the wind. Changing variables of this algorithmic wind, such as the direction and force, changes dynamically the effects on the modeled trees.
Virtual reality (VR) reguires a set of integrated hardware and software to create immersive 3D environments, where the user's field of vision is taken up by a mathematically rendered image that changes with the users movements and interactions. These environments can simulate the physical world, or be a completely invented world with its own rules. Another type of VR technology is "augmented reality" or AR, which can integrate 3D simulations into the real world. With goggles or a headset, a user can walk around a virtual object sitting on the actual floor in front of them. VR and AR are used in many types of immersive entertainment, especially games, as well as in special types of training that need to simultate difficult or dangerous situations, such as medical or military training.
Digital artists also utilize the computer's powers of simulation. They create new worlds for aesthetic pleasure, for storytelling or for perceptual disorentation. The works below only the scratch the surface in demonstrating how digital artists have worked in 3D animation, VR and AR.
The idea of 3D animation begins with modeling objects and mapping their position in a three-dimensional space. In the above early Pixar film, a 3D animation of a real hand gives a glimpse into how today's computer animation simulates fictional characters based on the gestures and movements of real-life performers.
In A Silva Guide to Birds from a Parallel Future, digital artist Rick Silva creates computer simulations that investigate the mechanisms of animation in the natural world. Unlike more realistic computer simulations, that try to smooth out movements in order to appear more "analogue", Silva's work exaggerates the digital, the on and off synchronization of nested parts, in the same way that an impressionist painter exaggerates the dance of light and shadow in a landscape.
Queerskins: a love story, by Illya Szilak
Illya Szilak's Queerskins is an example of character-driven immersive storytelling that makes use of integrated VR software and hardware: Unity Game Engine, Depthkit volumetric video, panoramic photography, 360 video and spatial audio.
The Thing Tableau by Mez Breeze is a text narrative mapped onto a 3D/VR sculpture. Suggesting the twilight state of insomnia, the work is designed for disorienting interaction by navigating around a very large robotic creature and clicking on text fragments.
Remix
Digital code makes it very easy to copy and paste. These are probably the most used tools in a personal computer. Code is also easy to manipluate and modify. Such "appropriation" or "collage" has a long history in the visual, audio and literary arts. Remix as a term emerges as music became digital in the 1980s and samples of tracks found their way into popular music, namely hip-hop. Creative appropriation and remix is now common across vernacular digital creativity and expression. YouTube alone is a vast repository of remix culture, where mash-ups of audiovisual media proliferate.
The Grey Album by Danger Mouse, is a well-known digital remix of the Beatles' White Album and Jay Z's Black Album. Because of obvious copyright violations, the album could only be aquired through p2p sharing technologies such as bittorrent. This YouTube video, a music video of a song on The Grey Album, is itself a visual remix of archival footage
Glitch and Noise
A "glitch" is a sudden surge of irregularity in behavior. Glitch art comes from the deliberate or accidental corruption of the organized digital code that make up a media file. Glitch or datamoshing, in which the code of a media file is meddled with, can be seen as generative of digital noise, when code is distorted from its intended message as an image or sound. The glitch of image files are particularly striking as the information for colors is stored in discrete units that create bursts of rainbow color when separated from the correct simulation.
Artist/writer Mark Amerika, known for his glitched, remixed and otherwise misbehaving texts, created an online work about a fictional glitch artist in the 1990s. The work is presented as a museum catalog of glitched gif animations.
Nonlinearity, Randomnness and Noise
The experience of nonlinearity in works of digital art is a result of an artist exploiting the computer's ability for random access. In computer science, the term random access, or more generally direct access, is the ability to access any element of a sequence or collection in the same amount of time. Any point is equidistant from any other point because each point - data, pixel or file - has an address and there is no noticeable space to travel to get from one point to another.
The ability to randomly access any point in a system implies that there is no hierarchy in the way data is stored. Data becomes malleable, networked and open to any kind of structuring. Non-linear video editing software enables direct access to any frame in a clip, without having to fast-forward to reach it. A book, with an index or table of contents pointing to specific page numbers, is certainly designed for random access, but it is still a linear sequence that needs to be traversed. It wasn't until computers became very fast at random access, making it possible to quickly navigate nodes in a network that entirely new forms of nonlinear and multilinear expression could emerge.
Randomness and chance have always been incorporated into the creative processes of art-making. A painter allows for the random irregularities of a brush stroke. A carver works with the chance irregularities of wood grain. But what is randomness really? Material conditions determine the texture of a brush stroke and natural forces determine the grain of wood. Randomness and chance refer to events that are not humanly determined, that is something unpredicatable. How can a computer, that only takes specific instructions and that is utterly deterministic in its workings, be a machine for randomness and chance? Randomness within a closed system becomes elusive.
The problem comes down to simply generating a random number. Random number generating functions in computer languages (such as Math.random()) create what is called a “pseudo-random” rather than “true-random” values. Pseudo-random values are generated by a series of operations performed to produce a sequence of numbers that repeat with a long enough period to be effectively random or sufficiently unpredicatble. Pseudo-random numbers are simulations of randomness and are in fact entirely determined and therefore predictable. Random.org offers "real" random numbers from measuring atmospheric noise. However, for most artists the pseudo-random number algorithms typically used in computer programming languages are just what is needed to create apparent variability and unpredicatle change.
Perlin noise is a technique developed by Ken Perlin that improves randomness in computer graphic renderings of natural noise in elements such as fire, smoke and water. Perlin developed the technique for the movie Tron in the early 80s and in 1997 won an Academy Award for his discovery. It works by interpolating between random values to create smoother transitions than the numbers returned only from Math.random().
The creative possibilties of randomness and nonlinearity are best illustrated in works of digital literature. Language is granular and made for a certain amount of poetic disorderliness. Nick Montfort and Stephanie Strickland's The Sea and Spar Between is progammed to randomly select from the most popular words of both Herman Melville and Emily Dickinson and to dynamically form these words into legible if not poetic stanzas.
Net Art
Hypertext works presented on the web in the late 90s were often within the context of “net art.” Net Art was a loosely defined online community of artists, writers and coders who embraced low bandwidth connections speeds, the accessibility of HTML and JavaScript to create interactive multimedia works, the frontier freedoms of expression and copyright in the early web, and the intimacy of connecting to a readership/viewership without intermediary publishers. Net art today is any work that incorporates the internet in the process of generating the work. One of the early net artists, Olia Lialina, has tried to collect websites from the 90s as a kind of folk art where the "user" was the creator, not the consumer of high-end works of entertainment.
Jodi was an net artist collective that created linked websites made of indecipherable, but visually intricate code. A peak at the first page's source code reveals a secret image in ASCII.
Olia Lialina's net art Summer enacts or makes visible the network that the artwork is built upon. The animation of the artist on a swing is automated to display single frames of the sequence delivered from web domains around the world. In the screen grab below, observe the changing urls in the browser address bar as the animation swaps images.
Artist Jon Rafman is an internet photographer. He "screen-captures" photos from his wanderings on Google Street View and then places these photos on a wall in a gallery. Akin to Marcel Duchamp's ready-mades, Rafman's digital ready-mades require only his selection of the found images. Are these his photos or Google's?
Another artist, Emilio Vavarella, travels on Google Street View screen-capturing all the “wrong landscapes” he encounters before others can report the error and prompt the company to adjust the images. Google’s technical glitches become the criteria for found art.
Mobile & Locative Art
Location-based or locative media refers to digital media that relates to or augments the physical location of the user. Geolocation (geocoordinates detemined by the Global Positioning System (GPS)) can customize the media content or advertising sent to a mobile device. Locative or mobile art is the creative use these spatial tracking technologies to offer multimedia experiences or stories. Such works are often based on the exploration of a specifc location, but they can also integrate the users own location in a media experience.
9.3
Generative Systems and AI Art
Generative art uses rules, procedures, randomness, data, simulations, or computational systems to create works that can vary over time. Contemporary AI art extends this longer history by using machine learning models trained on large collections of images, sounds, texts, videos, and code. In both cases, the artist often designs a process rather than a single fixed object.
Generative and AI art raise important questions about authorship and collaboration. Who is the artist: the person who writes the code, the person who selects the output, the person who makes the dataset, the person who writes the prompt, or the machine that generates the variation? Digital artists often work across all of these roles, designing systems, choosing constraints, curating results, and editing what the system produces.
“The important thing is not whether the computer is creative, but whether it enables us to be.”
— Harold Cohen, creator of AARON
Generative and Recombinant Processes
Generative art employs rules, data, simulations, algorithms, randomness, or external inputs to create an autonomous or semi-autonomous process. The artist designs the system and then allows it to produce variation. Recombinant art works similarly by selecting, rearranging, and juxtaposing elements from a set of media fragments. These approaches connect digital art to earlier traditions of chance operations, collage, conceptual art, and procedural writing.
Sound, Code, and Performance
Creative coding is not limited to images. Sound artists and musicians also use code as an instrument. Live coding turns programming into performance: the audience can hear the music changing while watching the performer write and revise code in real time.
Live coder Sam Aaron created Sonic Pi, a code-based music creation and performance tool. Sonic Pi was designed for both education and performance, showing how programming can become a playful and expressive practice.
“A programming environment which has sufficient liveness, rapid feedback and tolerance of failure to support the live performance of music is an environment ripe for mining novel ideas...”
— Sam Aaron
AI as Collaborator
Artificial intelligence has become an important area of digital art because it allows artists to work with systems that generate images, language, sound, code, and moving images from prompts, datasets, and learned patterns. Earlier generative artists often wrote rules directly. AI artists often work with models that have learned patterns from large collections of cultural material.
This does not mean that the machine replaces the artist. In many AI artworks, the artist chooses the tools, writes prompts, builds or selects datasets, curates outputs, edits results, and frames the work for an audience. AI becomes a creative material: not a neutral assistant, but a system with its own limits, biases, surprises, and visual habits.
Artist Sougwen Chung explores human-machine collaboration through drawing performances with robotic systems trained on her own gestures. Her work is useful for thinking about AI not as an automatic replacement for the artist, but as a partner in a feedback loop of movement, mark-making, memory, and improvisation.
AI as Archive and Data Environment
Some AI artists work with large archives of images, recordings, or institutional collections. Refik Anadol creates immersive data environments that transform massive datasets into moving images, architectural projections, and sensory installations. These works show one possibility of AI art: not simply generating a single image, but making vast cultural or environmental archives perceptible.
AI Cinema
Generative AI is transforming filmmaking as well as photography. Early experiments focused on using AI to generate scripts or dialogue, revealing both the surprising abilities and obvious limitations of machine learning. More recent systems generate moving images directly from text prompts, photographs, or existing video, allowing artists to create dreamlike sequences, animate still images, and explore entirely new visual aesthetics. Rather than replacing filmmakers, these systems have become new creative tools whose strengths and weaknesses are still being discovered.
Sunspring (2016), directed by Oscar Sharp from a screenplay generated by Ross Goodwin's recurrent neural network, remains one of the earliest AI filmmaking experiments. Trained on hundreds of science-fiction screenplays, the system imitates familiar cinematic dialogue while repeatedly missing the dramatic logic that human audiences expect. Rather than hiding AI's failures, the film turns them into its artistic subject, revealing both the promise and the limitations of machine-generated storytelling.
More recent artists have shifted from generating screenplays to generating images themselves. In 2023, German photographer Boris Eldagsen attracted international attention after declining the Sony World Photography Awards when his AI-generated image PSEUDOMNESIA: The Electrician won in the Creative category. His work questions photography, authorship, memory, and truth in an era when synthetic images are increasingly indistinguishable from photographs. Your Silence Has Been Recorded extends these concerns into moving-image form, using AI-generated imagery to create unsettling cinematic sequences that blur the boundary between photography, film, memory, and dream.
AI as Critique
Other artists use AI to critique the systems behind machine learning. Trevor Paglen and Kate Crawford's ImageNet Roulette exposed how image datasets and classification systems can reproduce stereotypes, errors, and social biases. Jake Elwes has used machine learning to explore gender, identity, and the unstable categories imposed by AI systems. These works remind us that AI art is not only about producing beautiful images. It can also reveal how machines see, classify, exclude, and misrecognize.
9.4
The Future of Digital Art
Digital art has always evolved alongside new technologies. Photography transformed painting. Film transformed storytelling. Computers transformed images, sound, animation, and interaction. Today, artificial intelligence is reshaping creative practice once again. Rather than asking only whether AI will replace artists, many artists are asking how new computational systems change creativity, authorship, originality, labor, and artistic value.
Artists as Designers of Systems
Many contemporary artists no longer create every element of an artwork directly. Instead, they design systems that generate possibilities. An artist may write code, train a model, assemble a dataset, construct prompts, guide an AI system, curate hundreds of outputs, and carefully edit the final result. The artwork emerges from a dialogue between human intention and computational process.
This way of working has precedents in conceptual art, generative art, procedural art, electronic literature, and interactive media. AI expands these traditions by making computational collaboration available to many more artists, while also concentrating power in the platforms and companies that control many of the most advanced systems.
Questions of Authorship
Generative AI challenges traditional ideas of originality and authorship. If an image generator has learned from millions of artworks, where does influence end and copying begin? Who deserves credit: the person writing the prompt, the developers who built the model, the artists whose work helped train it, the people who labeled data, or all of these participants together?
These questions are now being debated in museums, courts, universities, and creative industries. New laws, licensing systems, professional practices, and disclosure norms continue to evolve alongside the technology. For artists, the key issue is not only whether AI-generated work can be art, but how artistic responsibility should be understood when the creative process is distributed across tools, datasets, institutions, and human decisions.
The Ethics of AI Art
AI-generated art raises questions that extend beyond aesthetics. Many image-generation systems are trained using enormous collections of images gathered from the internet, often without explicit permission from creators. Building and operating these systems also requires computing power, electricity, water, minerals, and human labor. Understanding AI art therefore means considering not only what the technology can create, but also the environmental, economic, and cultural systems that support it.
Artists are responding in many different ways. Some embrace AI as a creative partner. Some reject it. Others use AI critically by exposing hidden biases, questioning ownership, revealing labor conditions, or showing how machine vision classifies the world. AI art is therefore not a single style. It is a field of practice and debate.
Looking Forward
Digital art continues to expand alongside virtual reality, augmented reality, robotics, artificial intelligence, bioart, immersive environments, and networked culture. The future of digital art will not belong to machines alone. It will emerge through new forms of collaboration between human imagination and computational systems.
Throughout this textbook we have seen that digital technologies are not simply tools but systems that shape communication, creativity, knowledge, and culture. Digital artists often work at the forefront of these changes, using emerging technologies not only to create new forms of expression but also to question how those technologies transform the ways we see, think, remember, and imagine the future.
9.5
Unit Exercise: Curate, Transform, Reflect
This exercise asks you to create a small digital artwork while documenting the tools, sources, and decisions that shaped it. The goal is not simply to make an interesting image, but to think critically about appropriation, transformation, software, AI, authorship, and preservation.
Part 1. Choose Source Material
- Find a public-domain, Creative Commons, or self-created image, sound, text, or video clip.
- Record where the source came from and what rights or license apply to it.
- Briefly explain why you chose this material.
Part 2. Transform the Material
Create two versions of the work:
- Manual or software transformation: alter the source using Photoshop, Illustrator, GIMP, Audacity, Premiere, code, glitch, layering, masking, color change, remix, or another digital technique.
- Computational or AI-assisted transformation: use a generative tool, prompt, filter, script, algorithm, or AI system to produce a second interpretation of the same source material.
Part 3. Curate and Revise
- Compare the two results.
- Choose the stronger or more interesting version.
- Revise it yourself so that the final work reflects your own artistic judgment.
- Give the final work a title.
Reflection Questions
- What original media did you appropriate or transform?
- What decisions came from you, and what decisions came from the software or AI system?
- How did the tool shape the style, mood, or meaning of the result?
- Did the computational or AI-assisted version surprise you? Did it become more generic?
- What ethical questions arise from your use of source material?
- How should the work be credited or disclosed?
- How might this work be preserved for future audiences?
Your submission should include the original source, both transformed versions, your final revised artwork, a record of tools or prompts used, and a short reflection of 250–500 words.
9.6
Glossary
- algorithmic art: art created through a set of rules or procedures.
- artificial intelligence: computer systems capable of generating or analyzing media and information.
- authorship: the question of who is responsible for the creation, meaning, and ownership of a work.
- creative coding: the use of programming as an artistic medium.
- dataset: a collection of data used for analysis, training, or generation.
- datamoshing: the deliberate corruption of video compression data to create visual distortions.
- emulation: the use of newer software or hardware to imitate an older system so that obsolete digital works can still be experienced.
- generative art: art created through autonomous systems, rules, or algorithms.
- glitch: an unexpected error or disruption that produces aesthetic effects.
- locative media: media experiences tied to geographic location.
- model: in AI, a trained computational system that can generate or classify outputs based on patterns learned from data.
- net art: art created for and experienced through networked technologies.
- noise: random or unintended variation in a signal or system.
- nonlinearity: a structure that allows multiple pathways through information.
- prompt: an instruction or input given to an AI system to guide its output.
- random access: the ability to retrieve any piece of data directly without traversing a sequence.
- remix: the transformation or recombination of existing media.
- simulation: a computational model of a process, system, or environment.
- synthetic media: media generated wholly or partially through computational systems.
- training data: data used to teach a machine learning model to recognize or generate patterns.
9.7
Bibliography
“Ada Lovelace.” Wikipedia, 29 Aug. 2019. Wikipedia, https://en.wikipedia.org/w/index.php?title=Ada_Lovelace&oldid=913026966.
Greene, Rachel. Internet Art. Thames & Hudson, 2004.
Hayles, N. Katherine. How We Think: Digital Media and Contemporary Technogenesis. University of Chicago Press, 2012.
“John Whitney (Animator).” Wikipedia, 29 July 2019. Wikipedia, https://en.wikipedia.org/w/index.php?title=JohnWhitney(animator)&oldid=908429741.
Lister, Martin, et al. New Media: A Critical Introduction. 2 edition, Routledge, 2009.
Manovich, Lev. The Language of New Media. Reprint edition, The MIT Press, 2002.
Paul, Christiane. Digital Art. Third edition edition, Thames & Hudson, 2015.
Ryan, Marie-Laure, et al., editors. The Johns Hopkins Guide to Digital Media. Johns Hopkins University Press, 2014.
Untangling the Tale of Ada Lovelace—Stephen Wolfram Blog. https://blog.stephenwolfram.com/2015/12/untangling-the-tale-of-ada-lovelace/. Accessed 17 June 2019.
Quaranta, Domenico. Beyond New Media Art. Brescia University Press, 2013.
Reas, Casey, and Chandler McWilliams. Form+Code. Princeton Architectural Press, 2010.
Tribe, Mark, and Reena Jana. New Media Art. Taschen, 2006.
Zylinska, Joanna. AI Art: Machine Visions and Warped Dreams. Open Humanities Press, 2020.
Crawford, Kate, and Trevor Paglen. “Excavating AI: The Politics of Images in Machine Learning Training Sets.” 2019.
Crawford, Kate. Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale University Press, 2021.
Menkman, Rosa. The Glitch Moment(um). Institute of Network Cultures, 2011.
Molnár, Vera. Vera Molnar: Variations. Various exhibition catalogues.
Paglen, Trevor. Invisible Images (Your Pictures Are Looking at You). The New Inquiry, 2016.