AI-generated music floods streaming platforms with millions of fake tracks, siphoning over $1 billion annually from real musicians' royalty pools

Streaming platforms like Spotify and Deezer now receive tens of thousands of AI-generated tracks per day. Deezer estimated in April 2025 that 18% of daily uploads — roughly 20,000 tracks — are AI-generated. Fraudsters use AI song generators like Suno and Udio to mass-produce tracks, then stream each one just enough times to collect royalties without triggering fraud detection. The Michael Smith case, where a North Carolina man allegedly used AI to generate hundreds of thousands of songs and earn over $10 million in fraudulent royalties, is just one known example. This matters because streaming royalty pools are finite. Every dollar paid out to a fake AI track is a dollar taken from a real musician. Estimates suggest over $1 billion per year is being extracted from legitimate artists' royalty pools. For independent musicians who depend on streaming income to survive — many earning fractions of a cent per stream already — this dilution is existential. A bedroom artist who gets 50,000 genuine streams per month might see their per-stream rate drop because the total pool is being divided among millions of fraudulent tracks. This problem persists because the economics of streaming platforms incentivize volume over quality. Platforms earn from subscriptions regardless of what gets streamed, so they have weak incentives to aggressively filter AI slop. The DMCA's safe harbor protections shield platforms from liability as long as they respond to takedown notices, but there is no practical way for individual musicians to file DMCA claims against millions of anonymous AI-generated tracks. The detection technology exists but lags far behind the generation technology, and platforms are reluctant to invest heavily in filtering when AI-generated content increases their total catalog numbers — a metric they use to attract investors and subscribers.