Should AI music be banned? The strongest case on both sides

Who has already banned AI music?

Some have, most have not. The moves so far split into three camps: outright bans, chart exclusions and labeling.

Notice the words "on suspicion". That is the hard part of any ban: someone has to decide what counts, and detectors make mistakes. Bandcamp is small and curated enough to try it. A service with tens of millions of tracks would be making that call thousands of times a day.

• 80%agree fully AI music should be clearly labeled (Ipsos for Deezer, 9,000 adults, Oct 2025)Deezer/Ipsos
• 52%say fully AI songs should not sit in the same charts as human-made songsDeezer/Ipsos
• 11%think fully AI music should be treated equally on chartsDeezer/Ipsos
• 65%say copyrighted material should not be used to train AI music modelsDeezer/Ipsos

What would a ban look like in practice?

Messy. Someone has to decide what counts as AI music, and the industry's own July 2026 labeling standard needed two categories to do it: AI-Generated, where AI made the whole recording or its main parts such as the lead vocal, and AI-Assisted, where people made it with some AI-generated elements. A ban on the first category is defensible. A ban that catches the second would hit producers who use AI for stem separation, sound design or a backing vocal, which ARIA explicitly still allows.

Then someone has to enforce it. Deezer claims its detector misses about 2 in 1,000 fully AI tracks and wrongly flags fewer than 1 in 10,000 human ones. Those are good numbers, and Deezer still uses them to tag and demote tracks, not to delete them. At the scale of a big service, even a tiny error rate means human artists accused and removed. That is why most platforms chose labels and demotion over deletion.

What are platforms doing instead of a ban?

Quietly cutting off the money. Tidal stopped paying royalties on fully AI music in June 2026. Deezer strips fraudulent streams from royalties and said in July 2026 it will remove AI tracks used for fraud and AI tracks nobody has streamed for six months. Spotify only pays tracks that pass 1,000 streams a year, and said in October 2026 that tracks likely made with AI are 10 times more likely to fall below that line. None of these is a ban, but together they remove the incentive to flood services with filler.

What should be banned instead?

1. Voice clones without consent. Tennessee's ELVIS Act (March 2024) added voice to the personal rights the state protects, and Spotify bans unauthorized AI impersonation. Every country should have that floor.
2. Streaming fraud. Fake plays should earn nothing. Deezer already strips fraudulent streams from royalties and removes AI tracks used for fraud.
3. Passing AI off as human. Require a clear label when an artist's identity or lead vocal is AI-generated. The July 2026 industry standard and Spotify's AI Persona badge are a start.
4. Unlicensed training. This one belongs to the courts and legislators. Universal and Sony's second suit against Suno (September 2026) and GEMA's German win will shape it more than any platform rule.

Will AI music be banned?

Not broadly. The direction in 2026 is labeling and sorting, not prohibition. Warner and Universal now license AI music companies even as Universal and Sony keep suing Suno, Billboard chose to identify AI songs rather than exclude them, and Spotify keeps AI Personas out of recommendations instead of deleting them. Expect more chart exclusions like Sweden's and Australia's, more filters like Deezer's planned on/off switch, and court-imposed limits on how models are trained.

Is AI music illegal?

Listening to it is not, and releasing it generally is not either. What is being fought over is how the models were built: Universal and Sony are suing Suno, Sony is suing Udio, and German courts have ruled against both Suno and OpenAI in cases brought by GEMA. Separately, the U.S. Copyright Office said in January 2025 that prompts alone don't make a work copyrightable, so purely AI output has weak legal protection.

"AI music is not real music." Is that fair?

It is a fair feeling and a weak argument for a ban. Music has absorbed every new machine, from the drum machine to Auto-Tune, after first being told it wasn't real. The better test is the one listeners already apply: does it move you, and were you told the truth about it? We dig into the philosophy in is AI music real music? and the ethics in is it ethical to listen to AI music?

For the data behind this debate, see AI music statistics and is AI music a fad?. For the legal fights in detail, read is Suno legal?