What to teach now, what to protect, and what to reinvent.
Dimitry Panov is a cultural researcher, educator, and co-founder of the Digital Renaissance Music Institute, launching in Dubai in 2025. His path spans music journalism, museum curation, university lecturing, and founding the Glinka School of Contemporary Music (later Moscow Music School). At DRMI, he’s rethinking music education for the digital age—where AI is everywhere, but human creativity still leads.
We sat down with Dimitri for a candid KHAMSA Q&A on the intersection of AI, culture, and music education.
١. If AI could be a bandmate, would you let it play bass, sing lead, or just tune the instruments?
I’d let it play bass. Bass is the foundation, the rhythm section that requires stability, precision, and a “machine-like” groove. That’s exactly what AI could do best. But lead vocals or solos? Those are always about human emotion, nuance, and imperfection.
That’s where AI can only ever remain an assistant, not a replacement.
٢. When was the first time you thought, “AI could actually change music”? Was it exciting, scary, or both?
Probably my very first encounter with something like this was about ten years ago, when I heard Japanese Vocaloids for the first time. It wasn’t fully-fledged AI, but it left me with a similar feeling to what I experience today listening to music generated by Suno.
Later, there were other generative music projects like Mubert that didn’t yet use LLMs but still produced completely synthetic compositions.
And then came Suno. Honestly, all of that felt more like a sideshow or a novelty attraction than a true revolution, fascinating, but not frightening. What I realized back then, though, was that this kind of technology could soon become a powerful assistant for producers in creating their own music.
And let’s be real: long before LLMs, music production was already heavily mediated by technical tricks, samples, sound libraries, automation, and algorithms.
Average listeners, and often even professional musicians or engineers, can hear a soundtrack without realizing that a portion of the sound wasn’t played by an orchestra but was instead pulled from a library or generated by a synthesiser. What LLMs have really done is streamline this workflow for creators: less painstaking manual labour, more creative freedom. And frankly, I don’t believe in a full or overwhelming replacement of humans in the act of composing music. We listen to music, just like we read literature or look at paintings, to access another person’s inner world and consciousness, and ultimately to discover something deeper about ourselves. With something that is 100% synthetic, that need can’t really be met, because we are not machines ourselves.
That said, let’s see what happens. Either way, it won’t be boring. And in any case, we’ll have to live with it and figure it out!
٣. What’s one human quality in music that you believe AI will never replicate?
The sense of risk. A musician improvising on stage is always taking a risk, of failing, of breaking down, but that risk is exactly what creates magic in the moment. An algorithm will never know fear or courage, and therefore it can’t create the same kind of tension and catharsis that a human can.
٤. If you had to teach an AI to understand music culture, what’s the very first track you’d play it—and why?
That’s an interesting question. If I said The Well-Tempered Clavier by Bach, that would be almost too obvious, maybe even cliché, though perhaps that’s what should be done. But instead, I’d rather load in some kind of “anti-music,” something that would force the AI to really struggle to figure out why it’s still considered music, and what exactly separates music from just sound. For example, I might choose any track from Plux Quba by Nuno Canavarro.
٥. Do you see AI as competition for musicians, or more like a new instrument in the toolbox?
Definitely as an instrument. AI can only compete with musicians on a very superficial level, with quick solutions and clichés. But a real artist will always turn any instrument into something more than it was originally meant to be. That’s what happened with the electric guitar, with samplers, with synthesisers, and that’s what will happen with AI.
٦. What’s the strangest or most surprising way you’ve seen AI used in music so far?
Probably the fan-made “deepfake” collaborations, like Frank Sinatra “singing” hip hop or Tupac performing Billie Eilish.
It’s absurd, surreal, but in a way it’s also a kind of postmodern collage art, a pastiche.
٧. If Mozart, Beyoncé, and a music-generating AI sat in on one of your DRMI classes, who do you think would learn the most?
The AI. Mozart and Beyoncé already carry centuries of human experience and artistry in their blood and memory. The AI, on the other hand, would have to absorb cultural context, meaning, and values from scratch.
٨. As an educator, how do you stop students from leaning on AI too much—without making it taboo?
By making AI part of the process. The point is not to ban it but to teach students how to use it properly. For example, we look at an AI-generated result and then ask: what’s missing here? What’s too mechanical? And how can the student push it further to make it real art?
٩. What excites you most about launching DRMI in Dubai, especially in a place where so many music cultures meet?
Exactly that: the cultural mix. Dubai is a new kind of cultural port where Arab, Indian, African, Western, and dozens of other traditions meet every single day. It’s a chance to create new hybrids and bring global attention to cultures that have so far remained mostly local.
١٠. Finally, if your life had a soundtrack generated by AI, what would you want the opening track to sound like?
Something like a mix of The Fall, Brian Eno and Kanye West: minimalism, a touch of chaos, sarcasm, roughness, layered with atmosphere — but with real depth underneath, bold ideas, and larger-than-life ambition.
Music that immediately signals: this is going to be unpredictable, but definitely not boring.
