Spotify: 660,000 Fake AI Tracks for $8 Million Misappropriated

A Sophisticated Scheme to Deceive Streaming Platforms

An American orchestrated a massive fraud using artificial intelligence to generate hundreds of thousands of music tracks. These songs, streamed by armies of bots, allowed for the diversion of approximately $8 million from streaming platforms. The tracks were fake, just like the listeners, but the amounts stolen from artists were very real. This individual recently pleaded guilty, marking the first criminal case of streaming fraud to be handled by the U.S. justice system.

Michael Smith, 54, a resident of North Carolina, admitted to conspiracy to commit wire fraud in a federal court in New York. He faces a sentence of up to five years in prison, with sentencing scheduled for July 29, 2026.

A Fraud Organized on Multiple Fronts

Smith's plan was based on three main pillars: production, distribution, and concealment. As early as 2018, he partnered with the head of a company specializing in AI-generated music to create an immense catalog consisting of hundreds of thousands of tracks without human intervention.

For distribution, Smith set up thousands of fake accounts on platforms such as Spotify, Apple Music, Amazon Music, and YouTube Music. An internally developed software allowed these accounts to stream his tracks continuously, 24 hours a day.

Concealment was the key to his success. Instead of concentrating listens on a few tracks, Smith spread the streams across his entire catalog. This strategy kept the number of listens per track relatively low while achieving billions of total streams. By using VPNs to simulate connections from various locations, he managed to evade detection by the platforms for seven years, from 2017 to 2024.

The result of this fraud: over $8 million in diverted royalties. Smith agreed to return the full amount, which totals $8,091,843.64, according to court documents. Spotify clarified that its platform accounted for only about $60,000 of the total fraud.

The Challenges of Music Streaming in the Face of AI

This case is not just a simple instance of individual fraud. It highlights a structural weakness in the "pro rata" payment model of streaming platforms. This system operates like a communal pot, where each fraudulent listen reduces the revenue of true artists. By manipulating listens through bots, Smith was not directly stealing from platforms like Spotify, but from the musicians sharing the same pool of royalties.

Artificial intelligence exacerbates this problem exponentially. Before 2018, Smith used his own compositions, but the shift to automatic music generation allowed him to multiply his catalog by hundreds. Today, Deezer receives about 60,000 entirely AI-generated tracks every day and has developed a detection tool to address this. Spotify, for its part, has removed 75 million tracks in a year and implemented new labeling policies. However, the issue of fake streams extends beyond AI and affects the entire economic model of music streaming.