The study took 50 participants with a high level of music knowledge and had them listen to excerpts of classical music either generated by a human, an AI called Deep Learning, or one generated by a non-Deep Learning algorithm.
They were not told who created the music, just to rate the piece on six musical dimensions: stylistic success, aesthetic pleasure, repetition, melody, harmony, and rhythm.
Results of the study found that the human-generated music overall was rated higher than those developed by a computer. The study stated, “On analysis, the ratings for human-composed excerpts are significantly higher and stylistically more successful than those for any of the systems responsible for computer-generated excerpts.”
But replacing producers and composers with computers is not the only concern many have about AI-generated music.
The algorithm in many AI programs sometimes takes direct music from their training or input data, causing many to believe it could lead to the original creator not receiving credit for their work.