Cedrik Sixtus got tired of finding AI-generated tracks in his Spotify playlists. So he built Spotify AI Blocker, a third-party tool that labels and hides suspected AI content. It filters over 4,700 suspected AI artists using community lists and signals like high release volumes and AI-style cover art. "It is about choice," Sixtus told the BBC. "If you want to hear AI music or if you don't." Hundreds have downloaded it. He warns using it may violate Spotify's terms of service.
Spotify's official position is hands-off. The company launched a test feature in April showing how artists used AI, based entirely on self-reporting through labels and distributors. That's it. Nothing more. "Our priority is addressing harmful uses like spam and impersonation, rather than trying to filter music based on how it was made," a Spotify spokesperson said. The company argues AI in music exists on a spectrum, not a binary. Meanwhile, widely suspected AI acts like Sienna Rose and The Velvet Sundown get treated like any other artist.
Deezer took a different path. The smaller competitor tags albums with AI-generated tracks and keeps them out of its recommendations. It built its own detection technology, trained to spot statistical patterns in audio, and recently started selling it across the industry. A Deezer-Ipsos poll found 97% of listeners couldn't distinguish AI from human-made tracks in controlled tests. Apple Music has promised "transparency tags" and will eventually require disclosure, but critics point out voluntary systems won't work if artists fear stigma.
The arms race is real. Bob Sturm, who studies AI's disruption of music at KTH Royal Institute of Technology in Sweden, describes the cat-and-mouse game in blunt terms. Detectors train on existing AI outputs, but as tools like Suno and Udio improve, the detectors need constant retraining. Fake scholars have already demonstrated the speed at which these tools can generate content, illustrating the constant need for updated detection methods. False positives are a real risk. Label a human musician as AI and you've done damage.
David Hoffman, a Duke University professor studying AI music's impact on artist livelihoods, isn't buying the complexity argument. He sees it as a lobbying tactic. "We can't draw the line, and therefore we shouldn't do anything," is how he sums up the industry stance. His take: label fully AI-generated tracks now, figure out edge cases later.
That position has teeth. Spotify's reluctance looks less like nuance and more like hedging.