DeepBach is an algorithmic program that uses machine-learning to analyse Johann Sebastian Bach’s compositions, and then produces its own variations. The results are convincing enough that many people can't distinguish between DeepBach and the real thing.
A Technology Review blog notes: "The team asked more than 1,600 people to listen two different harmonies of the same melody. More than 400 of them were professional musicians or music students. When given a DeepBach-generated harmony, around half the voters judged that it was composed by Bach. Even when confronted with music composed by Bach himself, participants only judged that correctly 75 percent of the time."
DeepBach is one of many algorithmic composition programs that have emerged in recent years.
The proliferation of these and other algorithmic composition programs is a key indicator of where music is heading. Increasingly, music is written by - or in collaboration with - algorithms. The way it works is that you give the program the parameters to work within - instrumentation, tempo, genre, mood - and it decides on the actual notes, melodies and rhythms.
So far much of this algorithmically generated music has been used for incidental and background music. But as we move forward, AI will play a very large part in building the templates for songs and albums to maximise time and minimise human labour.