US3R - Colors
Foreword by C Z A R I N A
Seattle-based electronic artist, multi-instrumentalist and producer Kristian Alexander, aka US3R, has been hacking pop music by creating unconventional sounds and weaving them into dystopian lyrical themes. Described as a combination of Justice, Daft Punk, The Weeknd, Billie Eilish, and Disclosure, his music feels familiar yet counters conformities and breaks fresh grounds. He also has a striking approach to his visuals and live performances — particularly the most recent one at the Static Realms’ Synthamania festival, where US3R cloned himself several times to perform each instrument in his songs.
US3R’s recent album INFLUENCE was created as the “Black Mirror” of electronic music. The other-worldly, grungy concept album explores social commentary about modern love, acceptance, and belonging in the era of social media, where people’s worth is measured by likes and followers.
In early 2020, the INFLUENCE Tour saw US3R playing shows across North America, before the global pandemic forcefully shut down the live music industry. Ever since, musical innovator US3R has leveraged his background in tech. Recently, US3R decided to use AI technology to explore a new frontier of AI-generated music, giving “hacker pop” a whole new meaning. The result is in his latest single, “Colors” - a smooth, vibrant electro pop track that brings holographic and colorful spectrum of images to mind, leaving a sonic trail of Aurora Borealis. We asked US3R to write about this unique experience.
About ten years ago, I set off to write a piece of software that would perpetually improvise music according to a variety of input parameters. I could change the beat and tempo, the genre of the music, the backing chords and layered lead melodies. My vision for that project was to be able to generate royalty free music as a service. Back then, my toy software project seemed innocent. I had no idea what would happen in music just a few years later.
Over the last couple years, there has been an explosion in the pace of advancement in machine learning and other AI technologies. Now, apps predict our every move, ML generates the articles we read, and sophisticated predictive modeling helps us rapidly find new treatments for diseases. Naturally, the creative technologies in the world wanted to apply this technology to music. With projects like Google Magenta AI, researchers were able to generate a whole new Nirvana song by using data from past Nirvana lyrics, chords, and melodies to train the new models. The scientific and music communities were shocked and amazed by the rapid advancement of music AI.
But what will we do when AI generative music is used for less benign purposes? As we saw in 2021 with the #FreeBritney movement, we learned that teams of people are more than willing to exploit artists with their likeness, their brand, and their fan bases to create a product to sell, even when the artist themselves do not consent to it. Like something straight out of Black Mirror, I believe we are at the edge of a new horizon in music where computers are no longer just tools to produce music. Instead, we are rapidly coming to grips with a new reality where AI may be, at best, accompanying the songwriting process, and at worst, replacing the artist altogether.
I wrote Colors to plant a flag in the ground and get a conversation started about what it means to use AI in music. Specifically, in Colors, I used neural networks to generate sets of lyrics that I could use in a new song. Once those lyrics were solidified, I attempted to compose music and feelings around those lyrics. In other words, US3R was the accompanying artist, and the AI was the centerpiece. At times, the phrases used in the songs felt alien; they contained phrases and sequences of words that I would never say as a human being. But, resisting the urge to polish the lyrics too much, I embraced these weird moments as evidence of the uncanny valley in the music - weirdness that is itself a byproduct of the robot - it was something unique that the AI was bringing to the table that I could not.
