This blog collects and documents artifacts arising from the contemporary collision of folk music traditions and statistical machine learning né Machine Folk. Each artifact reveals itself to the world through computational procedures numbering in the billions, and then someone or something breathing life into it to give it context. This “poetic research” is supported by ERC-2019-COG No. 864189 MUSAiC: Music at the Frontiers of Artificial Creativity and Criticism.
Some believe in a purity or divinity of Art, and the superiority of humans in making Art. Involving a machine in Art can be seen as a direct challenge to this purity, not to mention a demotion of the human to machine. This argument elevates the tangible product over the intangible experience, and submits to a rather narrow and mystical notion of creating Art — a human activity, full stop. For music, it is an activity that occurs between our ears in our sensations, thoughts and memories. Art is not contained on the page or in the frame. It does not stop when we leave the concert hall or the museum. Art is continuously steeped in being human. The human decision to involve technology in that activity is only that: the suspension of pigment in egg yolk to make it stick to a surface; the suspension of pigment in slow-drying oil to make it blend and layer in ways superior to egg tempura; the suspension of pigment in plastic medium to make it sculptable and fast drying; the use of a coarse horse hair brush to paint several strands of hair at once; the use of a palette knife to make sharp straight edges; the use of the principles of geometry and perspective to create trompe-l’œil. How is involving Ai in creating music any different? Computers have been able to beat the best human chess players for decades, but for some reason people are still playing and studying chess. Why? Because it is not all about the final state of the game.
One pitfall important to avoid when it comes to discussing Ai in the Arts is this: thinking that the terms “intelligence” and “learning” mean what they commonly mean when it comes to people. These are “suitcase” words with several meanings that can be confused, and must be unpacked. In the context of Ai, “intelligence” is a quite brittle thing that bears very little resemblance to human intelligence. And “learning” is merely the prosaic computation of numbers in an algorithm that represent probabilistic relationships in data. Falling into this trap can give power to the Ai that it does not actually possess, leading to dystopian fantasies. To claim Ai will destroy music is the same as claiming, e.g., photography will destroy painting, perspective drawing will destroy narrative, and so on. Yes, the future is filled with composers and painters, photographers and writers . . . The only real threats to the Arts are the lack of value for it in a society, and proscriptions of practice due to pressures economic, political, religious, etc.
Tunes from the Ai Frontiers 