Streaming platforms face a mounting wave of automated uploads designed to game royalty systems. Understanding what these moderation figures show requires looking closely at how services like Spotify handle algorithmic recommendations and fraudulent listening.
Filtering Recommendations
Spotify announced that it cut more than 50 million AI slop and spam tracks out of its recommendations this year. While these songs remain on the service, they are kept out of reach of the algorithm that decides what plays next. This figure is separate from the 75 million tracks the company deleted outright for serious or repeated violations, a count first shared in September 2025.
Listeners serve as the first line of detection for mass uploads, duplicates, SEO-stuffed titles, spammy playlist networks, and artificially short songs. Tracks that people skipped, ignored, or flagged consistently turned out to be slop or spam. Spotify layers this listener behavior on top of upload patterns, metadata, and third-party AI detection tools to identify likely spam with greater confidence.
What the Figures Do and Do Not Show
Reported figures highlight the scale of automated uploads, but they leave several questions unanswered. On peak days in June, more than half of what was uploaded to Deezer was fully AI-generated, yet those tracks pulled just 1% to 3% of the platform’s streams. Furthermore, as much as 85% of their 2025 plays were flagged as fraudulent.
Much of this listening is not human. Uploaders generate songs in vast quantities and loop them with bot plays to collect royalties. On Spotify, each track needs 1,000 streams a year before it earns a cent, a threshold set in 2024.
Despite sharing large numbers, companies often omit key details. Spotify gave no specific number for how much complaints about recommendations have dropped since changes began about six months ago. It also did not disclose what share of its uploads are AI-generated, whether the 50 million and 75 million figures overlap, or which specific detection tools it uses.
Industry Approaches
Different platforms are taking varied stances on AI-generated audio. Tidal took a hard line by refusing to pay out on fully AI-generated tracks over the summer. Deezer tags AI audio and plans to let listeners switch it off. Meanwhile, Spotify displays an AI Persona badge on AI-generated acts, has its Artist Profile Protection in beta, and is building its own AI remix and cover tool. Despite the influx of automated uploads, Spotify's more than 300,000 verified human artists still account for over 99% of what people search for on the app.