In the article “Creative AI: Computer composers are changing how music is made” by Richard Moss, current trends in music composed by computer programs is examined. Artificial Intelligence is an area of research in Computer Science that deals with creating computer programs that imitate (to some degree) the human mind’s learning process. Moss presents several examples of different kinds of computer music composition including: programs that are able to mimic a specific composer’s style (Well Programmed Clavier by David Cope), those based on biology (Iamus), and even one that automatically composes music for video games based on the live situation in the game itself (Mezzo). Many people wonder if A.I. composition can truly be creative, and if it is, can it put living composers out of work? On the latter, I don’t believe that A.I. composition will likely be able to supplant actual human composers. The question of a computer program’s capacity for creativity, however, is a much more complicated and subjective problem to address.
Composer Job Security
As a composer I can truly relate to the difficulty involved in finding work. Therefore, I also understand some composers’ reservations with research in the area of computer-composed music. The first argument that these programs pose no danger to composers is that they are largely created by composers or experts on music theory. None of the examples above could have been created without the input of composers because, much like a human composer, the computer composer must have a basic knowledge of music to know what it is doing. The “Well Programmed Clavier” was created by David Cope, who is a working composer and computer programmer. The program mentioned above that composes music for video games (Mezzo) also was created as a Doctoral dissertation in Music at UC Santa Cruz. It could be argued that computer generated music is just another tool in a composer’s toolbox. Now one might wonder: “What if a video game or movie studio purchases this software with the intent of using it instead of real people?” I will concede that this is a possibility, but I also firmly believe that the only companies that would do this have no intention of hiring a real composer anyway. If the A.I. music composer did not exist, they would probably just purchase music from a licensing company such as ASCAP. As a composer and software developer, I actually am embracing computer-generated music as an additional method for me to create new music.
The subject of a computer being “creative” is much more complex and subjective. How can anyone judge music in a scientific manner? The study of music aesthetics, the attempt to understand music’s effect on the listener, has been around since the 18th century. And if there is one consensus, it is that beauty is in the eye of the beholder. Some describe computer-generated music as dull, lifeless, and without “soul”. On the other hand, there are many instances where people are unable to discern music composed from a computer from that created by a human. As for my personal opinion, I believe that if a well-made program were to output a notated page of music for a human to play on their instrument, and that musician gave a practiced and thoughtful performance of that piece, then very few would know it was created by a machine. That being said, one should also consider that…for some… part of the experience of music is relating it to the composer’s state of mind when they created the music.
Despite all of the worries that Artificial Intelligence research inspires, I don’t think machine-composed music should necessarily be one of them. If anything, it is just another source of new and unique music that might otherwise have never existed. I would find it extremely interesting to hear what computers might be able to compose given different rule sets, techniques, and learning. While I personally doubt any computer program will be the next Beethoven, I would not be surprised if A.I. someday composed a piece of music that is loved by humans as much as one of Beethoven’s works.