By Ryan Leon, Feb. 25, 2025
Musicians use their talent and knowledge of music to create original melodies that have the power to enrich the lives of others, but with the accelerated growth of artificial intelligence, music can now be created without the use of any instrument.
Currently, there are multiple AI generators that assist musicians with creating music. Some examples include AIVA, Loudly, Amper, Beatoven and LANDR.
AI creates music by being fed preexisting data from other musicians and uses mathematical functions to represent that data, according to CPP student and Associated Students Inc. lead audio visual technician Nathan Brown.
“They’ll be represented as vectors, which is like if you have a point on a graph that’s in 2D space, a vector is in 3D or even more dimensional space,” Brown said. “They will go through this data set that you’ve given, take all of those things and turn them into numbers.”
The AI will then assign a name to those numbers. Using recordings of ‘90s music by Daft Punk, as an example, Brown also said the AI model will take every piece of those recordings and make them appear as a bunch of numbers, while another part of the model will go through and take those numbers and attach words to them such as “drums of Daft Punk from the ‘90s.”
After those two steps are completed, if given the command, the AI model will then take samples from those sets of words and produce a music sample based on those specific Daft Punk recordings.
While AI has become more popular over the last few years, the first example of AI-generated music was made between 1955 and 1956 when composers Lejaren Hiller and Leonard Isaacson created the Illiac Suite, a piece composed through a computer known as the ILLIAC I. Hiller and Isaacson fed composer Arnold Schoenberg’s entire library of music to the machine, according to Cal Poly Pomona music professor Sarah Wallin Huff.
“It made choices,” Wallin Huff said. “It made its own music based on the decisions it saw that Schoenberg had made, so it sounded like a student of Schoenberg had written it.”
Although the music was produced by a real human string quartet, it was a technological achievement to have a computer generate music by following the rules Schoenberg implemented.
Wallin Huff also shared a personal experience she had with a current, classical AI generator named AIVA. She witnessed a project for a TED Talk where they fed AIVA all of composer John Williams’ work.
Wallin Huff was hired as part of an orchestra that recorded the pieces being portrayed by AIVA, and she said it sounded like a student of Williams had recorded them. Wallin Huff added the music produced by the AI didn’t sound exactly like Willams but was eerily similar.
Yunsheng Wang, an assistant professor in the Department of Computer Science at CPP, said AI is the future. Wang said he feels AI will help humans in many fields, including musicians.
“I think for the artist, they still would like to have their creation, their music, their art,” Wang said. “But I think AI can assist them, not replace them.”
Brown himself dabbles in the AI music-making process. Though it’s only a hobby for him, Brown uses AI to assist him with producing beats.
One software Brown uses is called Ableton Live, a digital audio workstation that can be used to create beats. Brown showcased a recent track he’s been working on and said he used AI to assist him with reversing a sample he used. Through the use of a stem-separation program, Brown was able to separate the instrumentals from the vocals.
When it comes to things with big budgets, corporate owners will have no problem replacing human artists with AI produced music, but there’s still hope, according to Wallin Huff.
“Will (AI) replace live concerts?” Wallin Huff asked. “Will it replace live artistry? Will it replace experimental recordings? No. As far as the artistry is concerned, I don’t think that’s in any danger.”
Feature image by Connor Lālea Hampton
