Suno, one of the best-known AI music generators, says it will start watermarking the songs its users create.
According to The Verge, the company announced plans to implement new watermarking technology alongside a changed download policy, both aimed at limiting the spread of spammy AI tracks and increasing transparency about what is machine-made. The announcement came in a lengthy blog post from the company.
Engadget frames the change simply: Suno is adding audio watermarks so AI-generated songs are more easily identifiable. A watermark in this context is a signal embedded in the audio itself — something software can detect even after a track has been uploaded, re-encoded, and passed around, and which a casual listener would not notice.
Techmeme's summary of Terrence O'Brien's reporting for The Verge adds more detail on the scope of the plan: beyond watermarking, Suno intends to adopt fingerprinting technology and transparency tools, in addition to the new download policy.
The download policy is the part with teeth. Watermarks only help after a track exists; changing the terms under which users can pull finished audio out of Suno affects how much of it reaches streaming services in the first place.
Streaming platforms have struggled with a surge of low-effort AI uploads that clutter search results and siphon royalties from a fixed payout pool. If tracks arrive carrying a detectable marker of their origin, platforms gain a practical tool for filtering, labeling, or demoting them — rather than guessing.
It also puts a generator, not just a distributor, in the position of policing its own output.
This matters because the fight over AI music is shifting from whether the songs can be made to whether anyone can tell the difference — and Suno is now betting that the answer should be yes.