Spotify listeners who want to screen out AI-generated music still do not have a simple button to do it, even as frustration grows among some users and rival Deezer moves further toward labelling and filtering.
The issue has become more visible as generative AI tools can now produce polished songs with vocals, lyrics and instrumentation from short text prompts. The result is a fast-growing stream of tracks that many listeners may not be able to identify on their own, raising questions about transparency, artist payments and how much control users should have over what platforms recommend.
One frustrated listener, Leipzig-based software developer Cedrik Sixtus, built his own Spotify AI Blocker after finding his playlists increasingly filled with tracks he suspected were AI-generated. His tool, shared on code-sharing websites and downloaded by hundreds, draws on community tracking, release-pattern clues, cover-art signals and external detection tools to filter more than 4,700 suspected AI artists. Sixtus warns that using it “may violate Spotify’s terms of service.”
“It is about choice – if you want to hear AI music or if you don’t,” Sixtus told the BBC, saying he would prefer Spotify to label and filter AI-generated content itself.
Spotify is taking a narrower approach
Spotify has not ignored the pressure. In April, it began testing a credits feature showing how an artist used AI, but the system is voluntary and depends on what artists tell their label or distributor. Spotify said then that it was not a complete solution and that a comprehensive approach would require industry-wide alignment.
The company’s current position stops well short of actively identifying AI-generated music for listeners and letting them remove it from recommendations or playlists. A Spotify spokesperson told the BBC that its priority is addressing harmful AI uses such as spam and impersonation, rather than filtering music based on how it was made. The company also argues that AI use in music is not a simple yes-or-no category, but a spectrum.
That distinction is central to the debate. Some AI tools can generate a complete song from a prompt, while others assist with lyrics, arrangement or other parts of the creative process. Maya Ackerman, an AI and computational creativity expert at Santa Clara University and co-founder and CEO of WaveAI, said the call to label AI music can look straightforward from a distance but becomes complicated once partial or collaborative uses are considered.
Deezer is testing a tougher line
Deezer, a smaller Spotify competitor, has gone further. It has begun tagging albums that contain AI-generated tracks from services such as Suno and Udio and excluding those tracks from algorithmic recommendations or human-made playlists. Deezer says it uses in-house detection technology trained to identify statistical patterns in sound, and has recently started offering that technology to others in the industry.
Detection, however, remains contested. Bob Sturm, who studies AI’s disruption of music at KTH Royal Institute of Technology in Sweden, told the BBC that detection systems must keep adapting as AI music tools improve. False positives could also damage human artists wrongly marked as AI-generated.
Still, critics argue platforms can do more, at least for fully AI-generated tracks. A Deezer-Ipsos poll cited by the BBC found that 97% of listeners in a controlled test failed to correctly distinguish between AI-generated and human-made tracks, while about 80% of respondents said AI-generated music should be clearly labelled.
For now, the next pressure points are likely to come from industry standards and regulation. The music standards body DDEX is working on AI disclosure standards for music credits, and certain AI-generated content will require labelling under the EU AI Act from August 2026. Spotify has also announced features intended to highlight human artistry, including SongDNA and “About the Song” for premium users.
Until standards, enforcement and platform policy catch up, Spotify’s answer remains incremental: more credits and controls around abuse, but no listener-facing switch to remove AI music altogether.
Comments (0)