Study Suggests AI-Generated Music Is “Inferior” to Music Composed by Humans

AI algorithms haven't quite mastered the hit-making process just yet.

The AI robots are coming, but they may have hit a snag in their path towards becoming the music producers of the future.

According to a study conducted by researchers at the University of York, AI-generated music was found to be "inferior" compared to music produced by human composers. 

The study involved 50 participants with musical knowledge, who were instructed to listen to and rate audio excerpts based on six musical criteria. Some of the works were created by humans while others were generated by deep learning (a machine learning technique that teaches computers to simulate the behavior of the human brain) or non-deep learning algorithms.

The results indicated that participants were generally more favorable towards human-composed music, which received higher and more stylistically successful ratings than the computer-generated music.

Concerns surrounding the legality of AI-generated music have already been raised by many major copyright organizations, including the Recording Industry Association of America (RIAA). The study also raised concerns about the potential for ethical violations related to the propensity for deep learning tech to directly copy existing work.

"If Artist X uses an AI-generated excerpt, the algorithm that generates the excerpt may happen to copy a chunk of a song in the training (input) data by Artist Y," the study states. "Unwittingly, if Artist X releases their song, they are infringing the copyright of Artist Y."

The study goes on to suggest that the potential for infringement is "concerning" and suggests that certain regulatory requirements may be necessary to operate such algorithmic technologies.

You can find the full Machine Learning study here.