The initial coldness is striking. The darkness, the conceptual aesthetic line, the feeling of penetrating a complex universe, and I can't think of another adjective than cold. Not mysterious, not enigmatic, not unsettling, but cold. Metallic, squared. Raw. You immediately realize that they have chosen a very explanatory, very didactic, almost encyclopedic exhibition line.
Artificial Intelligence (AI) is presented to us as a great lesson, a very wide range of possibilities that seems to link very directly with the great exhibition on Big Data (Big Bang Data) that the CCCB itself offered us in 2014. Yes, in 2014; how time flies. But if that presentation was entertaining, artistic, aesthetic, colorful, and surprising in form, this exhibition ---co-produced by the CCCB and the Barcelona Supercomputing Center (BSC)---, which we can consider the second part of the story, is a cold and flat exhibition. It doesn't touch the skin or the soul. It touches information, yes. And, in that sense, it is an impeccable exhibition.
The first thing you discover is that AI is not the future, not even the present, but that it has been the past for many years now. It's been too long, by the way, that in digital art exhibitions we see the same anthropomorphic figures moving on a screen according to user movements: it even looks outdated. But when they connect Big Data with AI, which is a necessary and unavoidable connection, that's when you realize that we are much more in the hands of artificial brains, and for a longer time, than you thought: mobile phones, sensors, computers, and the internet accumulate millions of data that contribute to creating AI, and they do it without us realizing it.
The exhibition is full of installations that illustrate the great similarity between the creation of AI and the configuration of a neural networkIt is from these data banks that AI collects, creates its logics, builds images, and links concepts in a way that is entirely impossible for the human brain: and this is where it is as useful as it is dangerous. Facial recognition is an element already present in our security policies, as well as for ordering, for example, the driving of vehicles, but the problem with these utilities is the degree of dependence they can create. Indeed: What can happen if they fail? If AI fails in facial recognition even once, with a single person? Who is responsible for this and, above all, how to prevent it?
The exhibition is full of installations and screens, with detailed explanations on panels, which illustrate the great similarity between the creation of AI and the configuration of a neural network. Algorithms may seem arbitrary, but they are based on probabilities and logic, and can establish new languages that can even be moving: in the exhibition, you put your hand under a camera and it can take the form of a cloud, a galaxy, the sea or fire. With an almost unbelievable realism, so dynamic, detailed, vivid and live it is.
But we immediately move from these more artistic, more moving (few) interventions, to another succession of practical examples of AI: the help it provides in the early detection of forest fires, or in the identification of new planets, or in climate forecasting, or in the development of personalized medical treatments and surprisingly fast and effective simultaneous translation services. Other applications are more for entertainment, or curiosity, such as the ability to have all the phonemes and sub-phonemes of John Fitzgerald Kennedy's speeches to have a permanent JFK who can be asked to say anything. Or to highlight the reference to the Reprocessing piano + AI experience, presented at the Mies Van der Rohe Pavilion within the framework of this year's Sónar, where the machine, the contraption, reacts in real time to the music performed by a pianist: a curious musical duo for four hands is created, after all.
We also learn that in China, AI is already in charge of directing fire trucks, or regulating traffic light changes so that emergency vehicles can reach their destination with fewer interruptions, or so that traffic adapts to weather conditions or traffic density, reducing energy and light pollution. And then come the ethical questions, such as in whose hands these data end up, and in whose hands artificial intelligences are directed (we observe, however, that they can be in good hands and obey humanitarian objectives very efficiently).
And separately, a review of the prehistory of AI (the Bombe device aimed at deciphering encrypted German messages during World War II), or the first robots in the form of a mobile tarantula or a mechanized hand, or Kasparov's victories against the chess computer Deep Blue in 1996, with the charming final appearance of a state-of-the-art puppy that seems to respond to the stimuli posed by its artificial intelligence. Again, here a small oasis for sensibility: the installation Myriad (Tulips), by Anna Ridler, who photographed and classified more than 10,000 tulips to control the images generated by the neural network. The references to the golem or Frankenstein are less understood, if not to prick our imaginary of nightmares towards intelligent creatures that come to life, but placed in the midst of so much practical sense they seem a bit out of place.
The basic question is obligatory: is AI really artificial? What do we understand by intelligence?The basic question is obligatory: Is AI really artificial? What do we understand by intelligence? But the connection with the arts, music or literature is little exposed, which would have allowed for more sensitive and surprising spaces, even if reference is made to the inclusion of the DNA coding of Massive Attack members in one of their albums, or to being able to add synthetic DNA to the paint of a sprayor, however much it exists, the interactive installation where you can harmonize your voice in real time with Maria Arnal's.The conceptual connection made between AI and the DNA sequence is indeed interesting, since we always talk about decoded data, which has allowed lost smells to be resurrected (we can smell the essence of the flowers of a tree that disappeared in 1912 in Hawaii) or to redesign plants, mice, emails and (potentially) people.
At the exit of the exhibition, the tiled murals in the CCCB courtyard warn us, among their old flourishes: Del poc que tens, content seràs / si tot ho vols, tot ho perdràs. The taste that remains after leaving the exhibition is not one of danger, nor of the magnitude of what is coming upon us, nor even of an excessive (certainly) data load as if our head had to make a great final Big Bang Data. The aftertaste that remains, I fear, is that the exhibition could have been more fun, more sensitive, more artistic and more human. But don't listen to me. The brains that designed it must be of a superior intelligence, which I cannot access with my human measure. Undoubtedly, it must be that I am not prepared for so much artificiality.