A North Carolina man pleaded guilty to orchestrating a years-long music streaming fraud scheme that used artificial intelligence and thousands of bot accounts to siphon more than $8 million in royalties.

Michael Smith, 54, admitted to inflating streaming numbers for hundreds of thousands of AI-generated songs by deploying thousands of fake accounts across major platforms, including Amazon Music, Apple Music, Spotify and YouTube Music, according to court documents.

U.S. prosecutors said Smith worked with a co-conspirator and the chief executive of an artificial intelligence music company to acquire a vast catalog of computer-generated tracks, which he then uploaded to streaming services.

He used automated software to direct bot accounts to continuously play the songs, generating billions of streams between 2017 and 2024. To avoid detection, the activity was spread across thousands of tracks and routed through virtual private networks to mimic legitimate listeners.

“Although the songs and listeners were fake, the millions of dollars Smith stole were real,” U.S. Attorney for the Southern District of New York Jay Clayton said in a statement.

Smith created thousands of bot accounts — at times as many as 10,000 active at once — using fake email addresses purchased in bulk and outsourced labor to register the accounts.

Through the scheme, Smith collected more than $8 million in royalty payments that would otherwise have gone to legitimate artists and songwriters, prosecutors said. He also made false statements to streaming services, rights organizations and music distributors in an effort to conceal the fraud, according to court filings. Smith could face up to five years in prison.

Streaming platforms prohibit the artificial inflation of play counts through bots or other automated means.

The streaming service Deezer said earlier this year it is receiving more than 60,000 fully AI-generated tracks daily, prompting the company to expand its AI detection tools and consider licensing the technology across the industry.

Apple has also begun signaling a shift toward greater transparency, recently outlining plans to introduce metadata labels that disclose when and how AI is used in music production — a move aimed at helping platforms, distributors and listeners distinguish between human-created and synthetic content.