--- title: Artificial Intelligence, Artistic Future author: Geordee Naliyath date: 2026-10-18 url: https://geord.ee/2026/10/artificial-intelligence-artistic-future categories: - blog - thoughts --- # Artificial Intelligence, Artistic Future Most of us know "We Are the Champions" as something a stadium sings. Tens of thousands of people, a slow stomping beat, and a chorus simple enough to shout with your arms around a stranger. It does not sound like a song that needs explaining. Then I watched David Bennett's video, [We Are The Champions is more complex than you thought](https://www.youtube.com/watch?v=ZhIAUEUn1jM). It turns out it does. In music we usually look for the story in the lyrics. And that makes sense. For a long time music was laid over poems, and the words carried the narrative. But musical composition is a narrative skill of its own, and it reaches us by a different route. Major and minor chords, diminished and augmented ones, the tone of a guitar, the distortion, the rhythm and the groove. They make us feel something before we have understood a single word. The lyrics of We Are the Champions are about perseverance and triumph after hardship. The verses talk about paying dues, mistakes and criticism. Then the chorus turns into a victory chant, and a collective one at that. It does not say "I am the champion". It invites everyone around to share the win. And the music tells the same story, and arguably tells it better. The verses sit in a minor key, the scale we associate with dark and melancholy feelings. From there the song climbs through tension, withholding, achievement and celebration, moving from C minor towards F major along a path that is less predictable, though not unheard of. A lot of rock music of the era got by on a handful of simple chords. Mercury's writing has more chord changes, more inversions and moving bass lines, secondary dominants and borrowed chords - small surprises that the ear accepts. The struggle and the victory are written into the harmony, not just the words. Which brings me to 2026. We are in the middle of an AI revolution that touches every part of our lives, music included. And AI, for all its brilliance, is a great averaging machine. A model learns from millions of songs, and when it writes, it predicts what is most likely to come next. *Most likely* is another way of saying *typical*. Ask it for a song about triumph after hardship and it will do a good job: competent verses, a rousing chorus, a key change in the right place. I doubt it would take the path Queen took, or make us feel what Queen made us feel. > An averaging machine does not write anthems AI writes songs. Good ones, even. But to stand out, to be timeless, to step away from convention, that still takes a human. Now, someone will point out that we can turn up the temperature, or simply ask the model for secondary dominants and borrowed chords by name. True, unusual is easy. A random chord is unusual too. What Mercury did was deviate *on purpose*. The minor key holds the struggle, and the turn arrives exactly when the words reach the victory. It is deviation in service of a story, and that is a choice. The notes may well come from a model in future. The taste that decides which convention to break, and why, is still ours. This is not just about music. With AI accelerating production and build, making things is no longer the hard part. Work becomes a question of creativity and innovation. What makes the composition, or the product, stand out? What makes it a pleasure to use? And is that only about human joy and happiness? I do not think so. There is a productivity and performance side to it, at least for some products. Even in music, catchiness leads to more plays and more sales . A product that people enjoy is a product they come back to, recommend, and pay for. Delight shows up in the numbers eventually. So I see two kinds of work ahead: lift music and anthems. There will be plenty of lift music - run-of-the-mill products and productions. And that is fine. We need them where they serve a utility or stay in the background. We use them, put them aside, and forget about them. Much like the millions of songs that were produced and forgotten. Much like the millions of products listed on Amazon under alphabet-soup brand names that look like a password someone forgot to change. I wrote about [AI and overproduction](https://geord.ee/2026/06/ai-and-overproduction) a few months ago; lift music is where most of that overproduction will land, and hardly anyone will notice. Anthems are different. A composition, or a design, that people return to years later is not the most likely next note. It is the one nobody expected, and everybody recognises once they hear it. That still requires human creativity (or ingenuity, if you prefer). That is the artistic future of our artificially intelligent one. Let AI write the lift music, and it will do a fine job. Someone still has to write the anthem.