» As the future was predicted, the present became unpredictable, » says one of the narrators of This is the future, Hito Steyerl’s video installation exhibited at the 2019 Venice Biennale, whose images are all created by AI. Several narrators guide you through an artificial garden in continuous transformation where images are oscillating and blurry.
Flowers are born from nowhere and blend into each other, they sometimes leave room for bits of Venetian landscapes. The narrators highlight a characteristic point of this technology : everything you see is made up of past data that aims to anticipate the future, that is, to predict it. Your present here and now escapes you : warning ; whatever the network predicts may not be the case (…) let me make a prediction : none of this will ever happen” say the narrators. This garden is a potential and uncertain space that is out of step with a reality that you do not look at
» AI, the 21st century’s Pythia? « you think when leaving the space.
For some time now you have been seeing more and more artworks showing images like these : images that seem to appear and vanish at the same time. In the spring-summer 2019, two concurrent exhibitions explored the challenges of works created using artificial intelligence : Entangled Realities : Living with Artificial intelligence at the Haus der Elekronischen Kunst in Basel and AI : More Than Human at the Barbican Center in London (May-August 2019).
From the Palais de Tokyo in Paris in the Summer 2019, you remember images of cliffs taking shape to immediately collapse, an accelerated geology in motion that is at times almost surreal : images from Second Earth by artist Grégory Chatonsky.
Most of the artworks created by artificial intelligence use machine learning (the most widespread type of AI) that works with Generative Adversarial Network (GAN) , a system set up by researcher Ian Goodfellow in 2014 (for an understanding of how GAN works, see How dœs an artist work with AI?).
One of the first steps in working with GAN is to select corpuses of images that artists then transmit to the machine that they train to produce new types of images that are the synthesis in continuous motion of the former ones.
In particular, Gregory Chatonsky transmits to the AI machine various corpuses of aerial and geological images of the globe. The artist programs the machine to synthesize continuously image banks : the new images produced reveal another geological world.
In parallel with these two-dimensional images, which continuously synthesize those of the globe, Chatonsky creates sculptures with organic shapes from 3D files of living organisms and fossils. Like the 2D images of the globe in constant reconfiguration, they are syntheses of past and present « organisms » , « organisms that could have existed. It is a synthesis of the past and at the same time this synthesis of the past opens the way to another possibility, either that has not taken place or that could take place. » explains Gregory Chatonsky. We are in a state of permanent transformation. The AI machine builds other realities that are like the facets of a crystal.
By questioning the time to which AI images belong, artists who work with this technology throw us back to our present conditions of existence, for example by proposing a critique of our beliefs and dependencies in AI (This is the future by Hito Steyerl), or by echoing today’s ecological emergency (Second Earth ’s geology in accelerated motion by Gregory Chatonsky).
Works created with AI show you a present faceted of other realities : each time you find yourself in front of time spaces that show you a multiplicity of possible worlds. Artists reveal AI as a kind of crystal where realities multiply, intertwine, reverberate each other. These other realities refract into our present, they help us understand it, thinking it and anticipating it.
A new human resource
AI marks a turning point in the history of representation, as it has been the case with the successive inventions of photography and film in the 19th century.
Its developments will transform our understanding of reality and will possibly generate new human capacities, making us live in multiple artificial time spaces even more vertiginous than those we know with the web spaces. Will the human mind become more complex in order to live in these crystal time spaces? Will it gradually design machines that get even more complex?
The danger of AI remains that of any machine : the risk that the human mind loses the flexibility, reflexivity and adaptability that characterize it, and mimics the machine it produces.
Let us remember : only the human mind is disruptive in its ability to change the rules with which it lives, while the machine without consciousness can only execute programs. Thus, to the term « artificial intelligence » , some prefer to call the machine « augmented intelligence ». Gregory Chatonsky calls it « artificial imagination » in the rawest sense of artificial image production.