Artificial intelligence(AI) is seemingly everywhere nowadays. It warns swimmers for sharks , tracks students emotions and concentration levels and can even read minds. AI enables the creation of deep fakes, videos of real people seemingly doing things they never did, as well as disposes and amplifies the unjust, racial, chauvinist and other biases inherent in our culture and society. Besides these headline breaking news, AI is also complexly entangled in our daily lives for instance through our use of the internet, social media, or by just hearing the seasons hit songs generated or selected by some sort of artificial intelligence.
It might even seem that AI is capable of accomplishing anything and the progress being non-stoppable. Furthermore, industry and media further accentuate this notion. Therefore it might be beneficial to define the current AI systems and take a broader look at what is happening right now.
In general, AI systems can be categorised into three levels: weak, general and superintelligence. All of the current AI systems that are now in use could be broadly defined to be within narrow, or weak AI. This basically means that the AI systems are as competent as humans, or better, in one particular area, However, these systems are not capable of applying the same skills to another task or field. For instance, a discussion bot can be created to answer some helpdesk questions for a product or company, but if the AI is taken into other discourse (or are forced to face the witty of a creative user), the AI is completely lost. As such, we have to understand that in the context of AI, intelligence can be something completely different than we would comprehend it in everyday life.
The two other levels of AI: General AI and Super AI, differ from weak AI in that they are thought to have the very ability of justing and adapting to other tasks and situations. However, yet no real general (or Super in that matter) AI have been invented. However, media, some parts of industry and popular culture are heavily interested (as well as invested) in this part of AI, be it a humanlike supersystem, or malevolent computer destroying us and using us as batteries. The common lore is that when AI system is capable of adapting to other situations and improving itself, the system will do just that and get incrementally better and faster, finally reaching a technological singularity; A sort of Super AI that surpasses our own intelligence in many ways. Because of this, general and super AI fuel the wilder discussions of AI taking over the world or even giving us eternal life. But we have to comprehend that General and Super AI are still just hypotheses, not real things to worry about anytime soon.
As a philosophical side note, I have to mention how the whole AI discussion is propelled by a view that sees the mind as a separate from the body, and moreover how the concept of intelligence is transferred from embodied cognition to a formal manipulation of symbols (Hayles, 2008 p. XI) As such, the fantasies of downloading ourselves into computer systems to gain eternal life, seem even bit more far-off. The discussion of embodiment and AI should be crucial in our age but is seemingly nowhere. Dreyfus wrote already in 2001 that we should not forget our bodies in the age of digital technologies: ‘… without our bodies, as Nietzsche saw, we would literally be nothing. As Nietzsche has Zarathustra say: “I want to speak to the despisers of the body. I would not have them learn and teach differently, but merely say farewell to their own bodies — and thus become silent.”’ (Dreyfus, 2008 p.143–144).
The art world has also been affected by the AI hype. Last fall media-sites reported the first “AI-generated” artwork to be sold at Christie’s auction for a hefty 425 000$ price tag. Although, the artwork, a portrait of “Edmond de Bellamy” in 16th-century style, was actually created by Parisian art group Obvious, and their use of machine learning algorithms, the news fuelled an extensive discussion on the future of art. Some even concluded that we don’t need artists anymore. But does AI mean the end of human-made art? Or could AI provide a whole new dimension to art? How should we discuss AI and art?
Artists have for long been quick to experiment, use and misuse new technologies. Same holds true for AI. For decades now artists have created autonomous painter-robots as well as sought inspiration from AI. The use of AI is so broad that they could be broadly categorised in five categories. The first one sees AI as a material for art making: Besides creating portrait paintings, AI produces a wide variety of images. Most of these images are nothing like the portraits, but for instance, weird, faint and distorted human forms with multiple hands or mouths. Thing is that the AI systems do not see anything is wrong with them, as displayed by Janelle Shane, who in funny ways, show how AI gets things wrong and how that can be used as an art-making material. (See her keynote on Eyeo 2018 for example). Such play with material produced by AI defines the first category.
The second category focuses on using AI as a tool for art making. This means that just as we use brushes and paints to create paintings, we can use AI as a tool to express, relate, communicate something important to us. As in the case of an American artists Trevor Paglen’s work Adversially Evolved Hallucinations where AI is used to highlight the problematics within AI. The third category sees AI as the artist itself. Even though artists have for long created autonomous art-making machines, like Harold Cohens Aaron, a painting machine Cohen started developing as early as the late 1960s, few actually are considered to be artists, but more like the extensions of the artists themselves. Cohen, who passed away in 2016, has jokingly remarked that he might be the first artist that can have exhibitions of new works even after his death.
It seems that artists do not, yet at least, consider that AI could make art autonomously. This view can be backed by studies that state that jobs requiring original thinking and creativity are one of the least threatened by AI takeover (see here for instance) But can AI work as inspiration or guide in our post-digital age? The fourth category sees AI as a muse. This means such artistic practices where AI learns the artist's style or thoughts and shows the artists the paths to go forward. Or the artist can create a machine that mimics the artist's style and make art together with the AI.
But not all artists working with AI actually use AI. The final category, artificial stupidity, defines such practices that comment AI or make use of our comprehension of AI systems, without using them at all. Artist can create machines that are seemingly clever but are based on simple programming. Or the machines can fail miserably and create humor and commentary of our contemporary life, such as in Simone Giertz’s shitty robots.
The possibilities within the artistic use of AI are broad and deeply interesting. Art definitely can use AI and is being influenced by AI. However, I personally feel that the most prominent contribution art can make in the current post-digital society is to give us the possibility to stop and re-evaluate our current and future directions. Art can make visible such things of AI and our life with AI, that could otherwise be invisible. As an example art can bring about dissensus within AI as for instance, Sederholm and McQuillan suggests. These interruptions are crucial, in order to have time to see and think what is the question if the answer is AI.
Dreyfus, H. L. (2008).On the internet. (second edition), Routledge, New York
Hayles, N. K. (2008). How We Became Posthuman. University of Chicago Press.