From Fast Food to Fast Art: Do We Care?
In late 2025, we’re living in a time where anyone with a laptop and a few prompts can release a debut single. That’s exactly what happened this week with Silas Hart, a young newcomer from England, whose track Believer is featured as part my latest Sounds from The Unknown: New Music Mondays releases.
But here’s the twist. Silas isn’t real. Not in the traditional sense. He was co-created by me and a suite of generative AI tools: ChatGPT for lyrics, MidJourney for his initial portrait, Freepik Spaces for consistent video creation, and Suno for the music itself. The entire track, persona, and rollout came together in under a day. The result? A synth-tinged power ballad with emotional resonance and polished production. All of it manufactured, yet strangely still affecting.
This brings us to the heart of it. We’ve been here before. In the 1980s, the hit-making engine of Stock Aitken Waterman pumped out chart-toppers with factory precision. The Monkees were designed for television before they were ever a band. Milli Vanilli famously mimed to someone else’s voice. And now? We have ABBA performing as holograms and Gorillaz avatars packing arenas. We’ve long accepted blurred lines between authenticity and performance.
This tension between the real and the constructed isn’t new. In The Greatest Showman, one of the most emotional musical moments comes when Loren Allred provides the vocals for a song lip-synced on screen by Rebecca Ferguson. The voice is real. The performance is shared. The impact remains. In The Fifth Element, the alien diva sings an operatic piece composed to be impossible. Inva Mula, a human soprano, recorded the fragments. Technology stitched them into something unearthly. Audiences didn’t question it. They felt it.
We’re already starting to see it happen again. In late 2023, The Beatles released Now and Then, a “new” track made possible by isolating and enhancing John Lennon’s vocal from a decades-old cassette using AI. It wasn’t synthetic, but it wouldn’t exist without machine learning pulling his voice clean from the noise. The result feels fragile, haunting, and deeply human, precisely because it’s both old and new at once.
Then there’s AIsis: The Lost Tapes Volume One, the unofficial Oasis-inspired album built by a group of musicians using AI to emulate Liam Gallagher’s voice. The band didn’t reunite. They weren’t even involved. But the tone, the sneer, the unmistakable shape of that era? It’s all there. Some fans loved it. Some didn’t care how it was made. It felt like something familiar, returned.
So what happens when AI skips the stitching altogether?
When there’s no human voice at all?
And we still feel something?
That’s the question I’m sitting with.
The rise of “fast art” feels a lot like fast food. It’s easy to make. It hits certain cravings. It looks right. But it’s different from the slow, imperfect, unpredictable texture of a live performance. And maybe that’s exactly why we’ll see a return to the real. Not because AI isn’t good enough, but because people will start to miss the mess of it all. The sweat. The silence before the chord change. The voice cracking on the high note.
Of course, even those “real” spaces are tangled in tech. Auto-tune smoothing out imperfections. Visuals timed down to the millisecond. Backing tracks tucked behind live vocals. We’re already half in, half out. AI didn’t create the blur. It just sharpened the contrast.
Maybe we need something like the old Parental Advisory stickers. Not as a warning, but as context. A subtle cue. This song was written by a model. That performance came from a prompt. No shame in it. Just a little transparency. Maybe it matters. Maybe it doesn’t.
Because when Silas Hart sings about heartbreak, some listeners feel it. They connect to the voice. They fill in the rest.
And that’s what matters, isn’t it?
If the emotion lands, does it matter where it came from?
