AI music isn’t the problem — we are
From multiple perspectives, 2025 has been a turning point for AI-generated music.
- Quantitatively, French streaming platform Deezer reported a vertiginous increase in the number of tracks uploaded — with AI-generated music going from roughly 10% in January to around 34% by November.
- Culturally, Spotify introduced The Velvet Sundown, a psych-folk project whose analog-sounding tracks convinced thousands before the project disclosed its AI origins. Recent tests showed that 97% of listeners are unable to distinguish an AI-generated track from a human-made one.
- Legally and commercially, the three majors (Universal, Warner, and Sony) were in court against AI music companies such as Suno and Udio at the beginning of the year. By year-end, historic deals were signed: UMG × Udio, WMG × Udio, and WMG × Suno. In just months, the industry moved from confrontation to normalization. The market is reconfiguring at high speed, and music licensing will likely be deeply affected.
This leads to the question that terrifies many musicians: will AI-generated music kill human music?
But maybe this is not the right question. And a better question might be: is AI about to simplify and homogenize music — or has commercial music been simplifying itself for the past 20 years?
What we’ve done to music over the last two decades:
- Songs have become significantly shorter.
- The “bridge” between verses and choruses has almost disappeared.
- Key changes in commercial music have nearly vanished, and the same chord progressions are now common in both verses and choruses.
- Arrangements have become standardized across international hits.
- Dynamic range has been flattened through decades of the “loudness war.”
We optimized. We calibrated. We smoothed. We formatted. All to adapt to new music consumption habits: platforms, infinite feeds, and shrinking attention spans.
The result? An ultra-predictable soundscape. A globally uniform musical aesthetic. A musical language increasingly easy to model.
AI did not simplify music. It learned from music we had already simplified.
As François Pachet — former Director of Sony CSL Paris and Spotify Creator Technology Labs, and a pioneer of algorithmic music — recently put it, “AI is a deserved curse. It sends us a signal that, as humans, we should try to be much more original.”
Catalan pop star Rosalía took that signal and ran with it. Her album Lux — arguably the most audacious global musical offering of 2025 — proves that human artistry still has a pulse. Blending experimentation, flamenco, classical music, pop, rap, bold arrangements, and sung in 13 different languages, Lux was described by Amsterdam-based music agency Sizzer as “a beacon of optimism in a music landscape increasingly dominated by AI-generated acts and viral trends.”
The future of music isn’t AI vs humans. It could be humans daring to out-imagine AI.