Michael Smith generated hundreds of thousands of fake AI songs, ran billions of bot streams, and pocketed royalties that belonged to real artists. On March 19, 2026, he pleaded guilty — the first criminal conviction of its kind in U.S. history.
Every time you stream a song on Spotify, Apple Music, or Amazon Music, a small fraction of your subscription fee flows into a shared royalty pool. That pool gets divided among artists based on their share of total streams. It is the financial backbone of modern music — imperfect, underpaying, and yet the primary income source for millions of musicians worldwide.
Michael Smith understood this system better than most. He understood it well enough to steal from it for seven years.
On March 19, 2026, Smith — a 54-year-old musician from Cornelius, North Carolina — pleaded guilty in federal court to conspiracy to commit wire fraud. He admitted to generating hundreds of thousands of songs using artificial intelligence and deploying thousands of bots to stream those songs billions of times across Spotify, Amazon Music, Apple Music, and YouTube Music. The scheme ran from 2017 to 2024. The take: over $8 million in royalties that should have gone to real artists.
It is the first criminal conviction involving AI-assisted music streaming fraud in U.S. history.
HOW IT WORKED
The mechanics of the fraud were, in hindsight, almost elegant in their simplicity. Streaming platforms pay based on streamshare — not a fixed rate per play, but a proportional slice of a giant monthly revenue pool. If your music accounts for 1% of all streams on Spotify in a given month, you receive 1% of that month's royalty pool. More streams means a bigger slice. Fake streams means stolen slices.
Smith began with his own music. According to a Rolling Stone investigation, he emailed himself a breakdown of the operation: 1,040 bot accounts, each streaming approximately 636 songs per day, totalling 661,440 fake plays every 24 hours. At roughly half a cent per stream, that translated to an estimated $3,307 per day — or over $1.2 million per year.
But human-made music has limits. To scale, Smith turned to AI.
Working with what prosecutors identified only as "CC-3" — a co-conspirator designation typically used for individuals who cooperated with investigators in exchange for amnesty — Smith began purchasing and uploading AI-generated songs by the hundreds of thousands. The Mechanical Licensing Collective (MLC) database lists over 201,000 works attributed to Smith. At its peak, he had as many as 10,000 active bot accounts running simultaneously. He used VPNs, generated random artist names, and bought bulk email addresses to evade detection.
To avoid triggering fraud alerts, he never piled all the streams onto a single track. Instead, he spread billions of plays across an enormous catalog of disposable AI content — the sonic equivalent of money laundering through volume.
BOOMY AND THE QUESTION OF COMPLICITY
The identity of Smith's AI music partner — labeled "CC-3" in court documents — was never officially named by prosecutors. Billboard and Digital Music News independently identified that Alex Mitchell, the CEO and founder of AI music company Boomy, is listed as co-writer on hundreds of the 201,000+ songs registered to Smith in the MLC and ASCAP/BMI databases.
Mitchell has not been charged with any crime. When Billboard first reported the connection in 2024, Mitchell said he was "shocked by the details in the recently filed indictment" and that Smith had "consistently represented himself as legitimate." A representative for Boomy declined to comment following the guilty plea.
In May 2023, well before Smith's arrest, Spotify suspended Boomy's ability to upload new tracks and removed some existing music, citing "a review of potentially anomalous activity." It was also around that time — March and April 2023 — that the MLC halted royalty payments to Smith and confronted him about artificial streams. The scheme, which had run for six years, was beginning to unravel.
WHO ACTUALLY CAUGHT HIM — AND WHO DIDN'T
Not all platforms performed equally. Spotify, Pandora, and Deezer detected the fraudulent streams relatively quickly and removed Smith's content before significant payouts occurred. Spotify has publicly stated that its platform accounted for only approximately $60,000 of the total fraud proceeds — a striking figure given that the overall take exceeded $8 million.
Apple Music, Amazon Music, YouTube Music, and Boomplay chose not to comment when contacted by journalists. None have publicly disclosed how much of Smith's fraudulent royalties flowed through their systems.
The investigation was ultimately brought to a close by the MLC, which detected the artificial streams in 2023 and alerted federal investigators. The FBI's Complex Frauds and Cybercrime Unit carried the case through to prosecution.
THE SENTENCE AND WHAT COMES NEXT
Smith pleaded guilty to a single count of conspiracy to commit wire fraud — a significant reduction from the original three-count indictment, which carried combined potential sentences of up to 60 years. He agreed to forfeit $8,091,843.64. He faces a maximum of five years in prison, with sentencing scheduled for July 29, 2026. Smith remains free on a $500,000 bond.
U.S. Attorney Jay Clayton was direct: "Although the songs and listeners were fake, the millions of dollars Smith stole was real. Millions of dollars in royalties that Smith diverted from real, deserving artists and rights holders."
THE SCALE OF THE PROBLEM
Smith's case is extreme, but it is not an anomaly. Streaming fraud has been a growing problem for years — and AI has dramatically lowered the barrier to entry.
According to the IFPI's Global Music Report published March 18, 2026, Deezer reported receiving over 60,000 fully AI-generated tracks per day in January 2026, with 85% of streams on AI-generated music in 2025 classified as fraudulent — a 70% increase year over year. Beatdapp, a streaming fraud detection company, estimates that at least 10% of all streams globally may be fraudulent. The industry-wide cost: approximately $2 billion per year.
In response, Apple Music doubled its fraud penalties in early 2026 — raising the maximum fine from 25% to 50% of fraudulent royalties — after detecting 2 billion fake streams on its platform in 2025 alone. Spotify has updated its policies. Deezer capped individual user stream contributions at 1,000 per month.
But platform-level enforcement is reactive. Smith's scheme ran for seven years before anyone stopped it.
WHY IT MATTERS
The streamshare model means every fraudulent stream is not just money stolen from a pool — it is money taken directly from every legitimate artist on every platform where the fraud occurs. When Smith's bots generated 661,000 fake plays per day, real musicians received proportionally smaller royalty checks. Not because they made less music or fewer people listened — but because a fraudster diluted the shared pool.
This is precisely the kind of invisible manipulation that makes popularity metrics unreliable as a measure of quality. A song can have millions of streams and zero real listeners. An artist can appear to be charting while their catalog is entirely AI-generated and bot-driven. The numbers look identical. The music is meaningless.
It is why blind rating matters. At VoteMyAI, we strip away the clout — the stream counts, the follower numbers, the playlist placements — and ask real listeners to engage with the audio itself. No metadata. No context. Just the music.
In a landscape where streaming numbers can be manufactured by the billions, that kind of ground-truth feedback is not a feature. It is a necessity.
A GLOBAL TREND
Smith's case did not emerge in isolation. In March 2024, a Danish man became the first person convicted of streaming fraud in Europe, sentenced to 24 months in prison after generating millions of fake streams. In Brazil, prosecutors charged a suspect with uploading over 400 fake tracks and manufacturing more than 28 million fraudulent plays.
What is changing is the scale that AI enables. Creating 200,000 songs used to be impossible. Now it takes a subscription and a willingness to commit wire fraud.
THE BOTTOM LINE
Michael Smith did not invent streaming fraud. But he demonstrated, with uncomfortable precision, how AI transforms a small-scale scam into an industrial operation. For seven years, he ran what amounted to a royalty extraction machine — one that quietly siphoned millions from artists who had no idea it was happening.
The guilty plea closes his chapter. The problem it represents is only getting started.
If you want to explore AI music rated on its actual sonic merit — not its stream count — VoteMyAI is where that conversation is happening. Over 1,000 tracks. More than 6,500 blind ratings. No bots required.
And if you create AI music yourself, ElevenLabs remains one of the most powerful platforms for AI voice and audio generation — built for creators who want their work to stand on its own merit.