With the rise of AI music generators such as Suno and Udio, streaming platforms like Spotify face an influx of AI-created music. This surge brings up questions about the definition of “real” music for listeners and artists alike. In response to these concerns, Spotify unveiled new guidelines on Thursday that address three primary issues associated with AI: low-quality content, impersonation of artists, and transparency in AI usage in music production.
Spotify’s global head of music product, Charlie Hellman, emphasized the company’s intention to safeguard genuine artists from spam and deception, ensuring that listeners do not feel misled. He indicated that while the platform wants to protect authenticity, it also supports artists who choose to incorporate AI in their work.
The company is collaborating with DDEX, a music standards organization, to establish a new metadata standard aimed at identifying AI’s role in the music creation process. This includes participation in sound production, such as generating vocals and instruments, as well as assistance in mixing and mastering, as explained by Sam Duboff, Spotify’s head of marketing and policy.
According to Duboff, 15 record labels and music distributors have agreed to implement the new AI disclosure standards. However, he mentioned that a specific timeline for the rollout has yet to be determined, as labels will need to adjust their information submission processes to Spotify.
The streaming service is also enhancing its approach to combat impersonation, which includes the unauthorized use of another artist’s voice—whether genuine or artificially generated. Duboff highlighted that this policy will cover instances such as unauthorized AI voice replicas and deepfakes.
In the coming weeks and months, Spotify plans to introduce a spam filter to detect users attempting to manipulate the system. Duboff pointed out common spamming tactics, such as uploading tracks just over 30 seconds to generate royalty-eligible streams or re-uploading tracks with slight metadata variations. Over the past year, Spotify has reportedly removed 75 million spam tracks from its platform.
The company has also addressed allegations regarding the use of AI-generated music in its playlists to avoid compensating artists. Duboff firmly rejected these claims, stating that Spotify does not produce music—AI-generated or otherwise—and that all tracks on the platform are created, owned, and uploaded by licensed third parties.
During the press briefing, he did not provide information on whether AI-generated music might appear on Spotify’s editorial playlists.
Later, Duboff commented to Technology News regarding playlist curation: “Our editors focus on selecting music that resonates with audiences. We find that tracks primarily generated by prompts see minimal engagement. The conspiracy theories suggesting that we are promoting AI music for financial gain are untrue—all music on Spotify is provided by licensed third parties; we do not create or own any music, and we pay royalties accordingly.”
