Beatport Bans Fully AI-Generated Tracks, a Quiet Win for Southeast Asia's DIY Labels

Beatport just started saying no. As of this month, the store most working DJs still buy their tracks from will reject anything ruled fully or mostly AI-generated before it ever reaches the shelf. Tracks made with AI, mixing plugins, mastering tools, stem separation, still get in. Tracks made by AI, start to finish, don't.

The Filter Just Got Real

Beatport CEO Matt Gralen laid out the line the store is now drawing: "There is a difference between a tool that assists human creation and a system that replaces it entirely. Beatport is built on the former." That's the whole policy in one sentence. Auto-tune, comping, an AI mastering chain, none of that trips the filter. A track a model wrote from a prompt does.

The detection side comes from an expanded deal with Beatdapp, the fraud-and-rights outfit Beatport already worked with. Beatdapp co-founder Andrew Batey says the goal is to "bring greater transparency to generative AI while protecting human creativity." In practice, tracks flagged as fully or majority AI-made get held at ingestion and their rightsholders notified directly, rather than quietly listed and left for someone to complain about later. Anything judged AI-assisted still goes up, just tagged, so curators and buyers can see it.

Why DJs Aren't Mourning It

Beatport's own numbers make the case for itself. In a survey it ran ahead of the policy, 60% of respondents said they simply wouldn't play a fully AI-generated track, and 77% cited a flat preference for supporting human creators. Only 8% were open to AI music outright, with another 13% saying they'd consider it if the actual artists and labels got paid for it, which most generative tools don't arrange for anyone.

Context helps explain the urgency. Deezer confirmed earlier this year that around half its daily uploads, some 90,000 tracks a day, are now AI-generated. Spotify started labelling AI artist profiles with an "AI persona" badge earlier in August. Beatport is a much smaller, more curated store than either, built on DJ pools and label relationships rather than open uploads, but the same wave was reaching its shelves. This is a specialist store moving to protect what makes it a specialist store.

The Southeast Asia Angle

Here's the part that doesn't show up in the trade press coverage out of London or LA. A lot of what fills Beatport's house and techno shelves in this region comes from exactly the kind of small, human-run label that a fully-generated flood would drown out. Singapore's Darker Than Wax, fifteen years into pressing jazz-inflected house from a handful of people who still answer their own DMs, sells its catalogue on Beatport under its own label page. So does Bangkok's Pissawong Records, the outfit sampling Buddhist chant tapes into house music, whose Beatport shelf sits one search away from whatever a generative model spits out under a made-up alias.

That's the actual stakes for labels this size: not a crisis averted, since nobody claims AI tracks were about to bury Beatport's charts overnight, but shelf space and search relevance that stays worth fighting for. A DJ digging for a Bandung disco edit or a Yogyakarta gqom cut shouldn't have to wade past a hundred prompt-generated knockoffs first. Beatport's policy doesn't fix that everywhere, Deezer and the open-upload platforms remain a different problem, but on the one store this scene's labels actually rely on for sync and DJ pool money, it draws the line before the flood arrives instead of after.

Nothing changes yet for producers already using AI tools the ordinary way, stem splitting, de-noising, mastering chains. That was never the target. The target was the fully synthetic track wearing a label's name, and as of this month, Beatport says it won't carry one.