Computational art

A canvas is no longer necessarily made of fabric, nor are the strings of a guitar essential for creating music. In today’s world, art can also emerge from a line of code, a predefined algorithm or a form of artificial intelligence (AI) capable of generating works in a matter of seconds. Computational art is transforming the way we conceive of artistic creation and is gaining prominence with constant technological innovations.

Its emergence has sparked an intense debate: does technological mediation alter the very essence of what we understand as art? Is it a natural extension of disciplines such as painting or photography, or are we facing a new paradigm that challenges what it means to be an artist? While some celebrate the democratisation and the infinite possibilities offered by technology, others see it as a threat to authorship and the human essence. The only thing that is clear is that computational art leads us to rethink fundamental questions about the new cultural landscape amid one of the most controversial cultural debates of our time. The debate is on.

What is computational art?

Digital, computational, generative and AI art: how do they differ?

Although they are related and may overlap, each concept has its own distinctive features:

• Digital art: encompasses works created or modified using digital tools, such as illustration, 3D animation or video art.
• Computational art: utilises code, algorithms and computational processes as an essential part of artistic creation.
• Generative art: is based on systems of rules or algorithms that produce works or variations with a certain degree of autonomy, without the need for AI.
• AI art uses artificial intelligence models to generate or transform images, text, sound and other forms of artistic expression.

 SEE INFOGRAPHIC: How does computational art manifest itself? [PDF]

Historical milestones: from algorithmic drawing to generative artificial intelligence

Computational art has a history stretching back decades. Long before the rise of generative artificial intelligence, artists and programmers were already experimenting with computers, algorithms and interactive systems to explore new forms of creation. Here are some of the key milestones:

• The 1950s to 1960s. The first algorithmic experiments in machine learning: artists and programmers began to use mathematical rules and computers to generate drawings, patterns and compositions.
• The 1970s to 1980s. Art as a system: programmes capable of producing works with a degree of autonomy are developed and new forms of interaction between people and machines are explored.
• 1990s to 2000s. Interactive and connected art: the expansion of the internet and digital technologies drives the development of networked works, interactive installations and participatory experiences.
• From the 2010s and, in particular, the 2020s. The emergence and expansion of generative AI: advances in machine learning and subsequently in generative models broaden the possibilities for creating images, music, text and videos using artificial intelligence.

Artificial intelligence as a tool for art

AI-based art can be considered a subtype of digital or computational art. This technology uses algorithms to learn from large datasets to generate some form of artistic output, such as text, images, music or videos. The collaborative process between humans and machines is once again fundamental.

Key features include the ability to generate infinite variations from a single idea, the speed with which results are produced and adaptability to different artistic languages – from digital painting to music and literature. This has enabled many artists to explore previously unthinkable avenues, turning AI into a far-reaching creative laboratory for any user, amid the controversy surrounding the artistic qualities of the work.

This phenomenon has taken hold thanks to accessible platforms such as DALL·E, Midjourney and Stable Diffusion, which have democratised the ability to create images from simple text descriptions known as prompts. Furthermore, some AI-generated artworks have gained prominence in the market: in 2018 the auction house Christie's sold the portrait Edmond de Belamy, created using algorithms, for more than $400,000, marking a milestone in the visibility of this discipline. More recently, organisations promoting the convergence of AI and art such as the World AI Art Association and certain international exhibitions have helped to raise the profile of this field as a new chapter in cultural history.

Robot artists as a form of computational art

As well as being able to create art through programmes or algorithms, there are physical machines or robots designed to produce works of art. These systems can combine AI software with robotic arms, various sensors and tools that enable them to paint or even sculpt with partial autonomy and under human supervision.

One of the classic examples is Ai-Da, considered the first humanoid artist robot capable of drawing, painting and sculpting. Since its creation in 2019, numerous projects have emerged through which works have been exhibited in art galleries. These experiments challenge the boundary between authorship and tool while paving the way for a conversation about the role of technology in human creativity.

 Human creativity versus artificial creativity: can computational models replace human authorship?

The debate over authorship in technology-generated art pits different schools of thought against one another. For some theorists, the intervention of programming redefines authenticity rather than eliminating it. As such, the work is considered to be the result of a collaborative process between humans and machines, such that the computer or AI would be merely another extension of traditional artistic tools such as painting.

However, other critics insist that the essence of art lies in the human experience, among other factors. For them, works generated by computational tools lack human emotion and the experiences that nourish the supposed artistic outcome. Furthermore, they highlight legal, ethical and economic challenges regarding authorship: to whom does the work belong, to the programmer, to the user employing the computational tool, or to the community itself, whose data fed the system? Organisations such as the World Intellectual Property Organisation (WIPO) have emphasised these dilemmas, arguing that technological progress must be balanced with the protection of creators' rights.

What do computational models offer the artist?

In computational art, therefore, technology expands creative possibilities, but the artist continues to make fundamental decisions throughout the process. Their role goes beyond merely executing a work:

• Conceiving the idea: defining the concept, intention and message of the work.
• Designing the rules: establishing the instructions and parameters that guide the creative process.
• Choosing tools and materials: selecting the most appropriate technologies, data and resources.
• Selecting the results: evaluating, modifying and choosing the proposals that best express their intention.
• Providing meaning and context: interpreting the work and situating it within an artistic, cultural or social discourse.

The artist, then, not only creates works, but also designs, directs and interprets the processes that make them possible.