Artificial Intelligence creates music that is catchy, familiar, but ultimately uninteresting.
In a recent article published two months ago, the topic of A.I.-generated music stirred up quite a conversation. It marked a pivotal moment for SoulTracks, prompting the creation of a policy on covering music created by artificial intelligence. Subsequently, approximately 10-20% of the music submitted for review was based on A.I.
Over the Christmas holiday, I delved deep into the world of A.I. music firsthand. I used a popular A.I. music generation program to breathe new life into some older songs my brother and I had composed and recorded years ago. Initially, the goal was to freshen up these amateur recordings for some entertainment. The results, however, were surprising.
The aforementioned muffled soft rock ballad that depicted a love gone wrong was transformed into a crisp jazz-infused track filled with horns, while the murky mid-tempo pop tune emerged as a dance floor filler complete with handclaps and orchestral breaks. While I found thrill in hearing these songs in a new light, the response from friends and family was uniform – they preferred the imperfect, heartfelt originals over the polished, mechanized versions.
These reactions foreshadowed the conclusions I reached as I delved further into A.I. music generation tools. By providing prompts regarding lyrical themes, musical styles, and instrumentation choices, I effortlessly created a suite of instrumental jazz fusion pieces and “hard listening” tracks spanning various genres.
What surfaced from the A.I. tool were undeniably catchy and instantly appealing songs. However, each composition invoked a sense of musical déjà vu, as if I had encountered them before despite their novelty. This phenomenon is emblematic of A.I.-generated music, offering listeners a sense of ephemeral comfort in the familiar.
A.I. music programs are trained on vast catalogs of existing musical works, encompassing a multitude of popular contemporary songs. Consequently, the generated music can convincingly emulate the style and essence of those tracks without replicating them outright. This ability to mimic beloved musical styles sets A.I.-generated songs apart, charming listeners upon first encounter.
Nevertheless, a startling revelation emerged – while these new songs held instantaneous appeal, they lacked enduring substance. Despite resonating with the right notes, the absence of depth rendered them monotonous after a few listens. Music is designed to evoke emotions, and over time, these ersatz R&B or jazz tunes failed to move me in the same way that genuine, heartfelt music does.
Though I’ve entertained concerns about A.I. supplanting human creativity in music creation, this experiment instilled a sense of hope in me. While A.I. may significantly impact certain music genres, particularly ambient or background music due to its cost-effectiveness, true musical innovation remains a realm exclusive to human creators. Music’s capacity to evoke genuine emotion and forge connections is a realm where humans still reign supreme, steering music towards new expressions and frontiers.
The advent of A.I. represents a transformative force in the music landscape. Nonetheless, musical innovation will persist as a human-driven endeavor, ensuring that genuine emotions and relatable experiences continue to propel music forward. As we navigate this new era, we may find that the evolution of music is not as earth-shattering as anticipated.