A North Carolina man just became the first person convicted in a federal AI-assisted music streaming fraud case, and the numbers are staggering: hundreds of thousands of fake songs, billions of fake streams, and over $10 million in stolen royalties.
Michael Smith, 54, of Cornelius, North Carolina, pled guilty on March 20 to one count of conspiracy to commit wire fraud. According to the U.S. Attorney's Office for the Southern District of New York, Smith used AI tools to generate a massive catalog of computer-generated tracks, then deployed up to 10,000 bot accounts simultaneously to stream those songs around the clock on Spotify, Apple Music, Amazon Music, and YouTube Music. The operation ran from 2017 through 2024.
The mechanics were straightforward but industrial in scale. Smith bulk-purchased email addresses and outsourced labor to create thousands of fake listener accounts. Automated software directed those bots to play his AI-generated songs continuously. To avoid triggering platform fraud detection, he spread the activity across thousands of individual tracks and routed traffic through VPNs. At peak operation, his 1,040 accounts were each streaming roughly 636 songs per day, generating an estimated 661,440 daily plays worth about $3,300 a day in royalties.
$8 Million Forfeiture
Smith didn't act alone. Court documents reference a co-conspirator and an executive at an AI music company who helped him acquire the generated tracks. He also made false statements to streaming services, rights organizations, and music distributors to keep the scheme hidden.
U.S. Attorney Jay Clayton put it bluntly: "Although the songs and listeners were fake, the millions of dollars Smith stole were real."
Smith has agreed to forfeit $8,091,843.64. He faces up to five years in prison, with sentencing scheduled for this summer before Judge Koeltl.
What This Signals for Streaming Platforms
Every dollar Smith collected was a dollar that didn't reach legitimate artists. Streaming royalty pools are zero-sum. When fake songs accumulate billions of plays, real musicians get a smaller slice of the same pot. This case puts a concrete number on a problem the music industry has been warning about since generative AI tools made it trivially cheap to produce passable audio.
The platforms themselves bear some responsibility here. A seven-year fraud spanning four major services raises obvious questions about detection capabilities. Spotify and others have started implementing anti-fraud measures, including minimum play thresholds before tracks earn royalties and better bot detection, but those changes came well after Smith's operation was already printing money.
This is the first federal prosecution of its kind, but it almost certainly won't be the last. The tools Smith used are more accessible and more capable now than when he started in 2017. The barrier to running this exact playbook has only gotten lower.