Generative music is developed by the use of mathematical algorithms, or by the interplay of asynchronous music loops. It is not precomposed, as the performer does not know what the entire piece will sound like as it evolves. Rather, the piece is set in motion from a starting point and direction chosen by the performer. Often, multiple performances of the same piece will end up very different. Even very small differences in the starting parameters will cause large changes as time goes on, a demonstration of the butterfly effect.

An early example of generative music is It's Gonna Rain by minimalist composer Steve Reich (1965). In this piece, overlapping tape loops of the spoken phrase "it's gonna rain" are played at slightly different speeds, creating phasing patterns.

Brian Eno, considered the founder of ambient music, has used generative techniques on many of his works, starting with Music for Airports (1978). His works, lectures, and interviews on the subject have done much to promote generative music in the avant-garde music community.

Many software programs are now available to create generative music, such as Karlheinz Essl's Lexikon-Sonate (http://www.essl.at/works/Lexikon-Sonate.html) (1992-2004).