Artificial Intelligence and Classical Music Composition

This thesis deals with artificial intelligence composing classical music and ways of evaluating its performance by listeners. The text provides the first overview of the conducted experiments based on the so-called Turing test and, set up on the analysis of primary sources, it suggests possible improvements in terms of methodology. In the end, we propose an alternative test, which rejects the philosophical implications of the Turing test and, in contrast to the original experiment, also provides room for a music-theoretical analysis of the generated works