Into The Music
Apple Music and Spotify both launched new AI-related features today, and together they represent the most concrete industry response yet to the question everyone has been avoiding: how do listeners know whether what they’re hearing was made by a human?
Apple Music introduced “Transparency Tags,” requiring labels and distributors to declare any AI-assisted audio, artwork, lyrics, or video before uploading. The tags will be visible to listeners on the platform. Spotify, for its part, launched “SongDNA” in beta for Premium users, a tool for exploring the collaborators, influences, and connections behind tracks, alongside “Artist Profile Protection,” which gives artists the ability to review and approve music released under their name before it appears on the platform.
The Apple approach is about disclosure at the point of upload. The Spotify approach is about verification and ownership. They’re addressing different parts of the same problem: a streaming ecosystem that has been flooded with AI-generated content, including fake releases under real artists’ names, and a listening public that has been given no reliable way to know what they’re actually hearing.
Whether these tools actually change anything depends on enforcement. A tagging requirement means nothing if labels can simply choose not to tag AI content and face no consequences. Verification tools are only as good as the infrastructure behind them. Both platforms have economic incentives that cut against being too aggressive here, since AI-generated content is cheap to produce and generates the same stream revenue as any other track.
That these announcements came on the same day, April 1, is either a coincidence or a sign that the industry has reached some critical threshold of pressure from artists, listeners, and legislators. Probably both.
The timing here is interesting , both platforms releasing transparency features on the same day suggests this was quietly negotiated industry-wide rather than organically developed. From a production standpoint, what I actually want to know is how the transparency layer distinguishes between AI-generated texture beds and human-played instrumentation, because that distinction matters far more than the average listener realizes. The technical implementation will tell us whether this is genuine disclosure or just a PR compliance exercise.
Both platforms launch on the same day, both with conveniently vague ‘transparency tools’ that don’t actually tell you which tracks are AI-generated at the point of discovery. That’s not transparency, that’s liability management. Real transparency would be an actual label before a song plays. But that would hurt the numbers, so instead we get a press release and a settings menu buried three taps deep. The labels have been doing this since they started counting streams as sales , redefine the terms, call it progress.
What strikes me about this conversation is how different it looks from outside the Western streaming bubble. The mbira, the marimba, the mbaqanga guitar , these traditions survived colonization, survived the erasure of the recording industry’s early decades, survived being told they were not commercial enough to exist. They survived because they were tied to living communities, to ceremonies, to specific human meaning that no algorithm could replicate or replace. I am not against new tools. But “transparency features” that tell you a track has AI involvement after you’ve already streamed it 50 times are not protecting music , they’re protecting the platforms’ revenue while gesturing at accountability. The chimurenga tradition is alive because Thomas Mapfumo could not be separated from his people. Eddie Dalton has no people. That’s the whole problem.
There’s something almost poetic about the fact that both Apple and Spotify chose the same day to do this, which tells you everything , this was clearly a negotiated industry response to regulatory pressure, not a genuine reckoning. In Israeli pop and Mizrahi music we’ve had our own version of this identity transparency problem for decades, where the question of what counts as “authentic” Arabic influence versus appropriation has never been cleanly resolved by the market. The platforms picking which sounds get labeled and which don’t will replicate those same politics. An AI transparency tool designed in California will reflect Californian assumptions about what music is and who it belongs to. I’d want to know who was in the room when they designed the labeling criteria.