Spotify says it has removed more than 50 million songs it considers “AI-spam” from its recommendation systems, expanding a wider effort to identify artificial-intelligence content and prevent attempts to manipulate the streaming platform.

The company announced the figure in an update issued Tuesday, Oct. 20. Those tracks remain separate from the more than 75 million songs Spotify says it has already taken off the service entirely.

To identify what it has also described as “AI-slop,” Spotify said it uses a combination of signals rather than a single test. Its approach examines upload patterns and metadata while also drawing on third-party AI-detection tools, listener engagement and user feedback.

The latest action marks another significant step in a campaign that has included policy changes, new profile labels and a verification system intended to distinguish human artists. At the same time, Spotify is developing its own AI remixing tool, showing that the company is not rejecting the technology outright.

A growing AI enforcement push

Spotify tightened its AI policies in September 2025, when it reported that 75 million AI songs had been removed from the platform.

During a press briefing at the time, Spotify head of music Charlie Hellman drew a distinction between responsible creative use and efforts to exploit the service.

“We’re not here to punish artists for using AI authentically and responsibly,” Hellman said. “We hope that artists’ use of AI production tools will enable them to be more creative than ever, create more content that excites fans and offer the best possible experience on Spotify.”

He added that the company was focused on “the bad actors who are gaming the system,” arguing that Spotify could benefit from the positive side of AI only if it aggressively guarded against its downsides.

That distinction is reflected in the company’s different levels of action. More than 75 million tracks have been deleted from Spotify, while another 50 million-plus have been excluded from recommendations. Separately, profiles carrying Spotify’s “AI Persona” badge are also barred from personalized recommendations.

Human and AI profiles get separate labels

Spotify introduced its “Verified by Spotify” badge in April as a way to identify human artists. According to Tuesday’s update, more than 300,000 artists have now been verified through the program, with independent artists accounting for 70 percent of that total.

The streamer followed that initiative in August by creating the “AI Persona” badge for AI artist profiles. The feature officially rolled out in mid-September and prevents artists carrying the label from appearing in the platform’s personalized recommendations.

Together, the badges create visible categories for human artists and AI profiles, while the recommendation restrictions add a distribution consequence for accounts labeled as AI personas.

Spotify is also building with AI

Even as Spotify steps up its efforts against AI-related abuse, the company is working on an AI remixing tool of its own.

Universal Music Group, Merlin and Kobalt were among the first labels and publishers to license music for the tool. Believe joined that group on Tuesday.

The project aligns with Hellman’s earlier comments that Spotify does not intend to penalize artists simply for using AI. Instead, the company’s stated approach separates what it considers authentic and responsible use from content or behavior designed to game its systems.

ShowBiz Take

The scale of Spotify’s figures makes the significance clear: the company is no longer treating AI content as one uniform category. Tracks can be removed outright, excluded from recommendation systems or connected to profiles carrying a specific AI label. Human artists, meanwhile, have a separate verification route.

That layered response matters because Spotify is simultaneously restricting some AI-related content and licensing music for an AI-powered creative tool. Its position, based on the measures announced so far, is not simply “AI or no AI.” The dividing line is how the technology is used, how profiles are presented and whether activity is viewed as an attempt to manipulate the platform.

With more than 50 million songs now sidelined from recommendations, over 75 million tracks removed from the service and more than 300,000 human artists verified, Spotify’s AI policies are already operating at substantial scale.