THIS TALK TAKES PLACE ON THE 24TH OF SEPTEMBER 19:00 ON LIVE NOW
What would be the equivalent of a Turing test for artistic AI (artificial intelligence)? The common answer is that to pass such a test, a computer has to generate artworks “indistinguishable” from the ones created by humans. However, such an answer does not take into account the fundamentally different conditions of art making in different periods. Today art, music, media, and design creation takes place in a digital environment – i.e. it involves use of software apps, services and resources. In my talk, I will describe many features of this environment and explain how they pose crucial questions for defining a Turing test for “artistic machines.” I will also explain that computers have already passed the artistic Turing test as early as 1966, and therefore this test is no longer relevant. Therefore, I will propose a possible harder test for artistic AI. I call it a “Molnár test” – named after Vera Molnár, one of the pioneers of computer art in the 1960s.