US case alleges bots streamed AI songs to steal $10M in royalties

In short: Prosecutors say a man used bots to repeatedly play AI-made songs so he could collect more than $10 million in music royalties.

What happened

US prosecutors allege Michael Smith ran a fraud scheme from 2017 to 2024 that targeted music streaming services like Spotify, Apple Music, Amazon Music, and YouTube Music. The claim is that he used software “bots” (automated accounts that act like fake listeners) to stream songs over and over.

According to the indictment, Smith worked with an unnamed AI music company and a promoter to generate hundreds of thousands of AI-created tracks. Many had strange titles like “Zygotes” and “Zymoplastic,” and made-up artist names such as “Calm Baseball” and “Calorie Screams.” Prosecutors say this helped the tracks blend into the huge volume of real music uploaded every day.

The scheme allegedly spread listening activity across tens of thousands of songs to avoid raising alarms, peaking at about 661,440 streams per day. At its height, prosecutors say it produced more than $1.2 million a year in royalty payments.

Why it matters

Streaming services pay out royalties based largely on how many times a track is played. If someone can fake plays at scale, it is like stuffing a tip jar with monopoly money and then taking real cash out. The money paid to fraudsters comes from the same pool that should go to legitimate artists and songwriters.

The case also highlights a growing problem as more AI-generated music is uploaded. Deezer has said that a large share of streams for fully AI-generated tracks can be fraudulent, and it has started using detection tools to flag songs that are 100% synthetic so they do not earn fraudulent royalties.

Source: Financial Times