How Creative Machines are Redefining Art
Why shouldn’t software create “art,” really? ~@whitneymcn
Can software create art?
Yes, but only because Machine Creativity is changing the definition of art.
Software Redefines Art.
Art used to be an explicitly human form of expression. Nature can be wonderful, but — sticking to the definition — it’s not art.
What art used to mean is
a skill at doing something (from the latin Ars).
Its meaning changed for the first time in the 17th century. Then art became a mean of self-expression of the artist who creates it, without a function:
“Art for art’s sake” (MGM’s ars gratia artis).
Being an expression of one’s self, art pertains to the subjective world. Hence, art has always been very personal because it relies on our own appreciation of its beauty and the emotions it triggers. Especially in view of its original definition, its relation to design is evident (“Design is how it works”).
The definition evolved over time to include other aspects, such as description or external and internal exploration.
Mandalas, for instance, are used in Buddhism as an aid to meditation. Fledging beauty with emotion largely left to the external observer.
On the outside, currents like Process Art and partially Dadaism have shown that beauty and emotion exist also in methods and processes of creation. Art in the methods elevated the tools and the machines to center stage, put them under the spotlight …but the human always remained in control.
Software Redefines Art
Now the machines are taking the control over a large part of the creative process. The human gives the structures and sets them in motion, coming back later to appreciate the results. It’s not really a novelty — Brian Eno thought composers were more like gardeners than architects — but now the exponential is really kicking in, and it’s blossoming everywhere.
The definition is thus changing because the artist loses control over most of the process. Machines do most of the creation, leading to the question Who is actually the artist?
It’s hard to let go or be doubted, but it’s not a matter of control. A slender artist can delegate the sculpting of a massive demonic god to machines and other humans. You can create unique works of art by designing them and have them hand-painted — selfie, oil on canvas …for less than $50.
It is hard to let go because it is difficult for humans to make art. It’s a struggle, it takes so many resources and years of dedication. Very naturally, art is often associated with passion.
Passion ˈpa-shən/ noun: a strong and barely controllable emotion.
From the Latin pati: to suffer.
On the other hand, Machine Learning systems make it look so easy that it’s unfair. You can go to the App Store and make a self-portrait in Van Gogh’s style. Or whatever artist you feel like right now. Little emotion involved.
We want robots to make the tedious work for us, not take away the fun stuff. Indeed, knowing that it’s so easy to draw in Picasso’s porto-cubist style, the fun tends to disappear because there’s little joy and satisfaction in its accomplishment. Even the public and the critics (i.e., your friends and relatives) may glance over after a first “Cool!” If it’s done too quickly, it’s not satisfying. Why? Partially because there’s no anticipation, no time to build up dopamine, no time to struggle or make sacrifice giving value to the effort. Walking with Eno: Would a plant grow without the gardener? Maybe, and even less likely will it grow to be so pretty or good.
Where is the joy? Where is the accomplishment? If there’s no emotion or other reward, there will be hardly any intent to create.
Being fine with robots taking over where there’s no accomplishment, let’s not forget that photography had a similar effect, and incurred into a similar resistance (to know more, see Blaise Aguera y Arcas excellent post). Suddendly you could create portraits in matter of hours instead of days! Blazes, all the joy hath gone! No struggle! Passion no more!
Then photography became a type of art of its own.
Photographers started to profit from the acceleration of the new technique of representation, and innovated in creating works that were beyond imagination. The mended plant showed new, gorgeous patterns.
This means that it is likely that Machine Creativity will also become a new category of art. Creative Machines allow us humans to go beyond our imagination, where there is plenty of room to innovate.
Software Redefines Art.
Additionally, like with photography, the artist remains behind the machine. Setting the scene, pulling the trigger, and refining the result. Even if the Creative Machine takes care of most of the process, it does not become the artist; Even if it is capable — by observing our reactions — of autonomously producing something that is beautiful and emotional.
It’s a human artist that creates the work of art.
Because Creative Machines can’t make it up from scratch — yet — and they have no own intent of creating works of art — yet. Without intent, it’s not art, it’s nature. A new definition will be established when an unsupervised learning system will intentionally create by transforming a structured conceptual space. Art will then generalise to
An intentional form of expression
that is appreciated for its beauty or emotional power.
This post was istigated by @whitneymcn @aweissman @ODMagicOD @naval. It originally appeared as a tweetstorm to answer @whitneymcn’s question ‘Why shouldn’t software create “art,” really? Feels less and less odd with every passing year.’ Revised and extended.