David Cope is a composer and emeritus professor of music at the University of California, Santa Cruz, known for Experiments in Musical Intelligence, a program he began in the early 1980s that composed new pieces in the style of established composers. The program is abbreviated EMI and pronounced “Emmy.”
Cope started the project while stuck on an opera commission. Facing composer’s block and a deadline, he set out to build software that could analyze a body of existing music and generate new works that matched its style without copying any single piece. EMI worked by extracting patterns from a composer’s catalog - characteristic phrases, harmonic moves, and structural habits - and recombining them into original compositions said to sound like Bach, Chopin, Mozart, and others.
EMI became a touchstone in debates about machine creativity. The cognitive scientist Douglas Hofstadter built lectures around it, playing a genuine classical piece alongside an EMI imitation and asking audiences to tell them apart - a musical Turing test that listeners frequently failed. The work raised early versions of questions now central to AI music: whether style can be reduced to pattern, and what it means for a machine to compose.
For business readers, Cope’s EMI is an early, well-documented case of generative AI in a creative domain, decades before neural networks made the same idea commercial.