What will be the art of the future?
Not long ago, even the most optimistic predictions about the prospects of automation said it was impossible to imagine, for example, a self-driving car. Navigating actual traffic conditions has long seemed too challenging to formalize as a computer program. But there is already a debate about the consequences for humanity of switching to automated transportation. In addition, even a mediocre cell phone game application can seriously compete with the world's strongest grandmasters. Automating loan request processing, writing texts like press releases or even copycat poetry, and forming new clothing collections for mass consumption - are already a reality. In the field of medical diagnostics, robotization is developing, and machines capable of cooking food, caring for the sick, etc., are being designed. However, there is a type of human activity that, by definition, is not amenable to automation. We are talking about creative activities and the ability to create works of art.
Is a robot capable of mastering creative methods
Since the twentieth century, it has been accepted that further use of that pattern is meaningless as soon as a practice is recognized in which an artist's work is created. The essence of an artist's creation is to make a work of art as unrecognizable as possible, playing on the boundary between what is already known and what is not. Thus, in the case of art, we are dealing with the activity of creating more and more difficult to recognize patterns and developing a more and more complex system for identifying them. Many experts, meanwhile, believe that his history and physiology determine the work of the contemporary artist; he is too unfree due to economic reasons and his connection to art history. But anyway, the ability to produce non-automated solutions, resulting in works of art with new patterns previously unrecognized by art history, is still recognized as one of the defining characteristics of the human being. And if a human robot can display the required level of creative freedom, it will equal or surpass humans at some point.
Robot creativity
The development of modern technology is going in all directions, and artificial intelligence is increasingly asserting its position on the art scene. To imagine contemporary art in recent years without the intervention of artificial intelligence is already tricky. And this is not so much a new future as a concrete present. Self-learning systems have long begun to test creativity. For example, in 1970, scientists designed an algorithm that could write prose texts - though they were still relatively meaningless at the time. Since then, neural networks have learned to draw pictures, compose music and poetry, and create movie scripts. The principle of all algorithms is similar: they analyze a vast array of works of art and "create" their own based on the resulting patterns: a painting, a musical composition, a novel, etc. The creativity of neural networks is gradually being institutionalized. For example, in 2016, for the first time, there was a competition for artwork created by robots. This year, the PIX18 algorithm, invented by Creative Machines Lab, won the grand prize of $40,000: it was praised for its good stroke and ability to generate works based on the photos at its disposal. The structure and brushwork, on the other hand, were deemed to be close to Van Gogh's style.
Perception of the works
However, there is another critical issue: novelty. It is also the criterion by which we evaluate artists' creations. If the algorithms don't sketch or process photos but, for example, paint abstract paintings, can they really create something new? Developers at the Artificial Intelligence and Art Lab at Rutgers University have tried to answer this question by creating the Generative Adversarial Network (GAN). Previously, the algorithm learned from the responses of a single discriminator: it analyzed pictures, drew its own, and checked the result. It produced images similar to those it had learned before. The team took the next step in developing the network and added a second discriminator, competing with the first. Now the neural network analyzes thousands of paintings and, based on such a large sample, generates a list of conditions upon which the created picture can be classified as a work of art. At the same time, a second discriminator makes a list of styles and checks the picture for similarity to them - conducts an operation of verification. A new picture is born when the image is recognized as a work of art, not identical to any pre-existing styles. In addition, neural networks are already capable of creating cartoons. Computer programs can line up the images they themselves have drawn into a video sequence. The system adapts unique solutions for the automated creation of collages, simulating brush strokes on a canvas. The software can act painting techniques using multi-core processors - each thread controls a different brush. It allows you to "mix" brushes in unpredictable combinations, resulting in a more believable effect. The process of drawing - for example, a portrait - starts with marking out the regions of interest: eyes, mouth, eyebrows, etc. The application uses the neighborhood-growing method to partition the image and justifies the borders for each "area." It then paints each segment. Considering the light and environmental conditions, it can "paint" with pencils, pastels, watercolors, and crayons. The number of such technical possibilities is constantly increasing. One robot managed to captivate an audience with a musical composition, so they thought a human wrote it. And a short novel written by a Japanese robot almost won a literary prize. It raises another critical question: the question of the consumer's perception of art. Is there a distinction between our perception of a human-created work and that "generated" by a robot? Today, resources have been developed recognizing who wrote a particular poem - a bot or a human. The answer is not always obvious. It is ambiguous territory. There are works on the resource written by robots - even though people have attributed them to human authorship. Accordingly, we can assume that these algorithms pass the Turing test for poetry. A computer must convince 30% of humans of its "humanity" to pass the test. And yet, not only can we still mistake what a bot has written for human work, but vice versa, we mistake the work of humans for that of robots. There is a mixing of levels, a new understanding of texts and meanings, where the line between illusion and authenticity is blurred as we are used to it.
Creativity is an emotional impact
Another question is related to the essence of artistic work: how it differs from copying and reproducing past experiences. The American psychologist Colin Martindale proposed an original theory of creativity. According to his research, the creator's primary goal is to evoke emotional excitement in the consumer. It can be achieved by various means: novelty, the complexity of ideas, intellectual challenge, ambiguity, and ambiguity of interpretations and messages. A society in which the level of excitement ceases to grow (or begins to wane) is degraded. Martindale distinguished two stages of the cognitive process. The primary function is undirected, irrational thinking like dreams or daydreaming. The secondary method is conscious and conceptual; it is the solution of concrete problems and the use of logic. He applied a similar lens to the creative process: conceptual consciousness can discern and think logically, but it cannot create or deduce something it did not know before, ex nihilo nihil fit - nothing comes from nothing. Primordial thinking can draw analogies, build chains of association, and compare, generating new combinations of mental elements. It produces the raw material that conceptual thinking can process. The GAN described above works on a similar principle - one neural network "distinguishes," and the other "compares and finds associations. The algorithm follows the theory of creativity, producing new canvases that provoke an emotional response in people.
Neural Networks Help the Creator
Art and technology have always intersected and nourished each other (just remember the Renaissance, the experiments of Leonardo and Michelangelo). New materials, approaches, and inventions have often allowed artists to create masterpieces and entire art forms. So, in addition to the autonomous "creation" of poems, paintings, and music, neural networks now assist scientists in doing innovative research. The development of the modern music industry is focused on classified patterns that help literally build a mathematical model of music and "program" the desired effect of listening to a composition. In collaboration with Crimson Technologies, an international research team from universities in Japan and Belgium has released a particular machine-learning device that can identify listeners' emotional states and generate fundamentally new content based on the information gathered. According to experts, machines and programs for creating songs depend directly on automatic composition systems; their predetermined and stored volume of ready music material allows composing only similar tracks. The developers of such programs want to provide "machines" with information about a person's emotional state. In their opinion, this should help increase the interactivity of the musical experience. Scientists have conducted an experiment during which the subjects listened to music in headphones with brain activity sensors. The combined EEG data was broadcast to a robotic composer. The result was a greater engagement and more intense emotional response of the listeners to certain music. Such emotionally-connected interfaces have potential, as they can be used in health care to motivate people to exercise more often or simply cheer them up and lift their spirits.
The Art of Seeing
In The Art of Seeing, John Berger (Berger) noted that vision is primary to language. Knowledge affects our appreciation. According to Berger, any image is simply one of many ways of seeing, but our perception of an image depends on which way of seeing we use. So, the discussion of the creativity of algorithms motivates us to think not only about how programs "create" but also about how we ourselves perceive creativity. Neural networks can write poems, and we sometimes confuse them with human ones, but our perception and reading give them meaning. For example, for an algorithm, words, strokes, colors, and sounds are just a set of signs that it can put together into a rhythmic structure. It is a raw material behind which the robot does not see the content, the field of meaning. At least, not yet. Robots cannot give meaning to objects, nor can they give global cultural value to works. An AI can create an ingenious symphony or a combination of rhymes organized correctly graphically. Still, only human recognition will allow all this to achieve the status many people desire - to really be art, not to seem like it.
Artificial Intelligence as an Artistic Tool
Like everything else about artificial intelligence, no one knows precisely how this technology will evolve in the future. However, more and more experts are putting forward theories about a new human-machine relationship focused on cooperation rather than the domination of one over the other. Many researchers see AI as a tool to help people create exciting works of art in the future. The creativity of machine intelligence can enhance the imagination: having better tools for creativity will allow more people to develop and artists to go much further than they could have gone on their own.
Using artificial intelligence in contemporary art
AI-based tools are already being used to automate time-consuming processes that previously needed to be done manually. And the results show not a potential encroachment of AI on the human artist's work but rather a benefit to creativity. Companies that create creative tools that have become the industry standard have been adding AI features to their innovative digital software in recent years. They hope it will speed up the workflow by automating its routine component, giving artists more time for self-expression and experimentation. AI integrates into software unobtrusively but has a surprisingly significant impact, from machine-learning tools that make finding specific video frames faster to features that let you color outline drawings at the touch of a button. The best AI features can help artists by freeing them from repetitive routines. This view is based on a Pfeiffer Consulting study in which a large portion of the creative professions expressed that they were not afraid of being replaced by AI and that they see the main potential of AI and machine learning as being applied to tedious, noncreative tasks. For example, this could mean an intelligent photo cropping feature that automatically recognizes the subject in the frame or automatic tagging to help people find stock photos faster. That said, artist control is still necessary. AI can't replace creative inspiration. Other AI-based features can significantly impact productivity in the creative process. One function, for example, is generating tags for video descriptions or selecting similar photos from the Internet. Other AI tools can find more severe uses in an artist's work, such as the automatic coloring tool created for comics and animation. With a bit of input from the artist, the program can automatically color a black-and-white outline image. AI-based coloring tools can play a significant role in the future of two-dimensional animation; they can give the artist the freedom to experiment by not having to spend time coloring every frame. Automating a large part of the process leaves more time for different ideas and learning other visual language options because they can be implemented much faster. AI is taught through sets of outline drawings made from color illustrations. The system is built on a deep learning algorithm that combines computer vision tools, such as those used in unmanned cars, with visual content creation systems. It is critical to note that artists retain copyright on uploaded and generated images, and the data will never be published. Developers are optimistic about how the tools they create will serve to benefit artists rather than trying to replace them. They believe that AI-based features are simply a type of tool within the capabilities of digital art, and creative people will use these tools in the best way they can. Computational creativity doesn't just live by painting. Machines can also compose music using an artificial intelligence system.
AI will affect the entire art market
Artificial intelligence is not a direct threat to artists. They can use it in their work - as long as the algorithms and input data are open to reproduction or created by them. However, in the long run, the advancement of AI art will have tangible consequences for the market. There is an opinion that with the arrival of new technology, non-AI art will be transformed. In the same way, the invention of photography once influenced the development of painting: it gave rise to Impressionism, Expressionism, and other schools interested in expressing emotions and unique human perception. According to many artists, AI will lead to novel-like forms and even unexpected and provocative conceptual works in painting. After all, such art is a direct rendering of description. Despite this, our perception of art is saturated with emotion. In this respect, it will be difficult for artificial intelligence to approach living authors. Thus, severe changes in the development of technical means have affected all spheres of human activity, including creativity. During the last decades, culture has been undergoing global changes due to the development of technology: computers and digital, due to which new artistic genres are being formed. Extreme creative possibilities have opened up in areas such as virtual reality, three-dimensional animation, the Internet, and interactive systems. AI is far from being a myth. Even though creativity, according to many, may exclude the introduction of such innovations, as in any sphere, AI has a place in culture. AI has enriched the arts with new tools and possibilities, and its rational use of artificial intelligence brings many benefits.