AI-made music is creeping into the charts. Nobody can agree on what to do about it.

Key points

• Fenix Flexin's single "Rubberz," released in June, reached No. 58 on the Billboard Hot 100 with more than 35 million Spotify streams.
• The track sounds nothing like the artist's known style, drawing widespread suspicion that AI generated key parts of it.
• Record labels are scrambling to set internal rules as AI music tools, software that can compose and sing in seconds, become mainstream.
• No US law currently requires artists or labels to disclose when AI made or heavily shaped a released song.
• The debate is no longer whether AI music exists on major platforms; it is what, if anything, anyone can do about it.

The song of the summer debate usually comes down to which track you cannot get out of your head. This year, it has turned into something stranger: a fight over whether the winning song is even real.

"Rubberz," a June single by rapper Fenix Flexin, sits at the centre of that fight. The track climbed to No. 58 on the Billboard Hot 100, a chart that ranks the most popular songs in the United States, and logged more than 35 million plays on Spotify. Those are real, significant numbers.

But the song itself sounds wrong in a way that is hard to pin down.

Fenix Flexin made his name in trap music, a hard-edged style of hip-hop, as one half of the rap duo Shoreline Mafia. "Rubberz" is nothing like that. It is a mildly dark, 80s-style synth-pop track featuring a lead voice that does not sound like him at all. Listeners compared it to Morrissey; others heard echoes of Men at Work's "Down Under." More than a few called it unsettling.

That uncanny, slightly-off quality is exactly what critics of AI-generated music point to when they try to explain what separates machine-made songs from human ones. Not obviously fake. Just not quite right.

Why does this matter to ordinary music fans?

If you stream music, pay for a subscription, or buy concert tickets, the rise of AI-generated tracks affects the money flowing to human artists. Streams translate directly into royalties. Every AI track that earns those streams instead of a human-made one shifts income away from working musicians.

The Guardian first reported on the wider industry anxiety surrounding cases like "Rubberz," noting that record labels are now scrambling to adapt internal policies as tools that can generate a full song, complete with vocals and instrumentation, in seconds become widely available to anyone with a laptop.

Musicians are vocal about the threat. The core complaint is not just artistic. It is economic: AI tools have almost no up-front cost, which means a label or an independent operator can flood platforms with content that competes directly with songs that took months and real money to make.

What rules exist right now?

Very few. No US federal law currently forces artists, labels, or distributors to label a song as AI-generated before it lands on streaming platforms. Spotify and Apple Music have policies against fully AI-generated tracks that disguise themselves as human artists, but enforcement relies largely on self-reporting.

Several bills are moving through Congress that would require disclosure or create new rights for artists whose voices are cloned without permission. As of publication, none has passed into law.

For listeners, the practical upshot is blunt: you have no reliable way to know, from a song's listing alone, how much of it a human actually made.

Common questions

Can I tell if a song I'm listening to was made by AI?

Not reliably. Detection tools exist but none is accurate enough to be definitive, and streaming platforms do not currently display AI-disclosure labels on individual tracks.

Does this affect the money artists earn?

Yes. Streaming royalties are paid per play, so AI tracks that accumulate streams draw from the same pool that pays human musicians, reducing what working artists earn overall.