Music that no one ever really listened to. Fans who never existed. And yet millions of dollars in the bank. It sounds like the plot of an absurd comedy, but this really happened.
Michael Smith, a 52-year-old man from Cornelius, North Carolina, pleaded guilty to fraud in federal court in New York in March. Using his method, he defrauded streaming platforms and thousands of real musicians of more than 8 million dollars. And all he needed was artificial intelligence, a few automated programs, and a considerable amount of audacity.
Music generator and AI bots
Smith did not need a band, a studio, or talent. Using AI tools, he created hundreds of thousands of tracks and uploaded them to platforms such as Spotify, Apple Music, Amazon Music, and YouTube Music. He then unleashed an army of bots that played these tracks day and night. The result? Up to 661,440 streams per day and an annual income of around 1,027,128 dollars.
The bots simulated the behavior of real listeners. At first, the platforms' systems were unable to detect them. And Smith happily collected the money. By his own estimates, he was earning approximately 3,300 dollars per day, or more than 1.2 million a year. This continued from 2017 until 2024.
An investigative report by Rolling Stone revealed a bizarre detail: Smith operated 1,040 accounts on streaming services, each playing around 636 tracks per day. He was not some technological genius working out of a garage. He was a "fortysomething suburban dad" who ran a network of urgent care clinics. To him, streaming fraud was simply... a side income.
Money from the pockets of real artists
Streaming platforms operate on the basis of a shared pool. The money is distributed proportionally according to the number of streams. Every fake stream of Smith's AI music therefore took a slice of the pie that should have gone to someone else. Someone who actually wrote, recorded, and released the track.
The words of U.S. Attorney Jay Clayton say it all: "The songs were fake. The listeners were fake. But the millions of dollars Smith stole were real." Think your favorite independent artist earns decent money from streaming? Think again. Then imagine those earnings being reduced even further by people like Smith.
Guilty plea and sentence
Smith was arrested in September 2024 and initially denied any wrongdoing. Eventually, however, he reached an agreement with prosecutors and pleaded guilty to one count of conspiracy to commit fraud. He will be sentenced on July 29, 2026.
He faces up to 5 years in prison and must repay exactly 8,091,843.64 dollars. The case is being handled by the Southern District of New York and represents one of the very first successful federal prosecutions for AI fraud in the music industry.
AI music is a problem
Smith's guilty plea may have brought a small measure of satisfaction, but the music world still faces a deeply unpleasant reality. Over the past year, Spotify removed 75 million spam tracks. The French service Deezer receives 60,000 fully AI-generated recordings every day. And according to Billboard, Suno, a company that enables music to be created from text, generates 7 million tracks per day.
According to Deezer, 97% of people cannot distinguish AI music from human-made music. That is a figure that should concern anyone who appreciates vibrant, authentic creativity.
Smith's case is not an isolated one. These are bots listening to bots. This entire absurd loop operates at the expense of those who create music of their own free will, with their own emotions and their own story.
Spotify promises to invest in detecting artificial streams, but the reality is more like an endless game of cat and mouse. On one side are platforms trying to reassure artists and the public. On the other are those who see the algorithm merely as a loophole they can crawl through with a briefcase full of fake notes. Smith may lose his freedom. But the problem he helped set in motion will never disappear.
Sources: musicradar.com and futurism.com
