Man jailed for 18 months after making $8m from AI music with 1,000 bots

What happened?

Michael Smith of Cornelius, North Carolina, uploaded hundreds of thousands of AI-generated songs to streaming platforms including Spotify, Apple Music and Amazon Music between 2017 and 2024. He then used more than 1,000 bots to artificially inflate the listening figures for his tracks to billions of plays, earning roughly $8 million through the platforms' per-stream payment schemes.

How he did it

Smith used 52 cloud service accounts, each with access to 20 bots, putting 1,040 AI agents at his disposal. Connected to VPNs, these bots could stream around 636 songs per day, amounting to roughly 661,440 daily plays. The US Department of Justice put the estimate even higher, saying Smith deployed up to 10,000 bots at a time.

The sentence

Smith was charged with wire fraud conspiracy, wire fraud and money laundering conspiracy in September 2024. He pleaded guilty to a single count of conspiring to commit wire fraud in March 2026. Facing up to five years, he was ultimately sentenced to 18 months in prison and ordered to forfeit $8,091,843.64. He will also face two years of supervised release after his sentence.

Google's side

The news comes as Google announced it has watermarked 240,000 years' worth of AI-generated audio and 180 billion AI images and videos using its SynthID Detector tool. Initially available to pro users, SynthID Detector is now open to anyone and works with content made by Google, OpenAI, Nvidia and Kakao, with Apple expected to follow soon.

Why it matters

Smith was not prosecuted for making AI music, but for how he monetized it. By flooding platforms with bot-led artificial engagement, he misrepresented the popularity of his tracks and reaped the financial rewards. According to the DoJ, streaming fraud diverts funds from musicians and songwriters whose songs were legitimately streamed by real consumers. Apple Music Vice President Oliver Schusser said more than a third of the tracks on the platform are "100% AI," yet they get "below 0.5% of usage."