A Deezer/Ipsos blind test of 9,000 people across eight countries found that 97% of listeners couldn't tell AI songs from human ones. So the sound problem is basically solved. Suno can generate something indistinguishable from a competent indie track in about forty seconds. That's genuinely impressive, and dismissing it with 'but it has no soul' is lazy criticism.
The problem is that 'sounds fine' was never really the whole job. Deezer's own platform data from 2025 makes the gap concrete: AI music accounts for 28% of daily uploads and roughly 0.5% of actual streams. You can flood the catalog. Capturing attention repeatedly, over time, from people who actually care about who made the thing. That's different.
The Velvet Sundown Problem
The Velvet Sundown hit a million monthly Spotify listeners in 2025 before anyone confirmed it was fully AI-generated. Once that confirmation came, the backlash was swift and the academic scrutiny followed. The reveal mattered. Audiences feel something different about artists they invest in, and 'investment' is the operative word. You follow an artist across albums, arguments, breakups, reinventions. There's nothing to follow when the entity doesn't exist.
Xania Monet, an AI gospel persona created using Suno by poet Telisha Jones, became the first AI artist to debut on a Billboard airplay chart and hit No. 1 on the R&B song sales chart. Remarkable as a technical and commercial milestone. Also immediately sparked pushback from artists like Kehlani. Chart position is a metric. It is not the same thing as legitimacy, or a relationship, or a reason for anyone to still be listening in three years.
That line is more damning than any industry critic could write. The platform that makes the tool is watching people optimize the wrong thing. A better prompt gets you a better song. It does not get you a person who texts their friend about you at midnight.
The Gap Is Disclosure, Not Quality
Here's what the data actually says about where the career ceiling lives: 62% of U.S. consumers say they're less likely to engage with AI music, and 80% want it labeled. Those numbers don't contradict the blind test results. They explain them. People can't hear the difference when they don't know to look for one. But when the origin is disclosed, the relationship changes. That's not irrational. When you love a record, you're partly loving the idea of another person having made it through something.
Suno's recent executive hires, including a new VP of Artist Marketing, suggest the company understands this. You don't hire for artist marketing if you think the product speaks for itself. The tool is mature. The surrounding apparatus of identity, trust, and continuity is not, and it may not be buildable through prompts at all.
None of this means AI music is worthless or that its role in the industry won't grow. It will. But the current framing, that better generation equals eventual parity with human artists, skips the part where audiences don't just consume music. They attach to the people behind it. Suno can hand you a song. It cannot hand you a story people want to keep following.