Is this the future of AI scamming? Michael Smith, a 52-year-old from Cornelius, North Carolina, allegedly used AI to automatically generate hundreds of thousands of songs that he then put on streaming services like Spotify, Apple Music, and Amazon Music. He’s accused of using an army of bots to “listen” to the tracks that he created with no effort and likely no musical talent.

That netted him over $10 million in streaming royalties.

Now, Smith has been indicted for what the US Attorney for the Southern District of New York and the FBI called a “brazen fraud scheme.” He has been charged with wire fraud conspiracy, wire fraud, and money laundering conspiracy.

The scheme dates back at least to 2018, when Smith partnered up with an AI music company and a music producer to mass-produce an enormous catalog of tunes. He created several fake accounts to spread the music around so it wouldn’t seem like one person suddenly could create fully produced tracks at rates faster than any known person in the history of the human race. He has a network of bots listening to them at a rate of around 661,440 streams per day.

Streaming services usually have fraud detection systems in place, but at every step of the way, Smith figured out a method of circumventing detection. For a while at least.

He pulled it off by creating randomized song and artist names that were convincing enough to provide an air of legitimacy. Garbage, total nonsense band names like Calm Identity and Zygotic Washstands and many thousands of others, according to the indictment.

Variety reports that a nonprofit group called the Mechanical Licensing Collective, which ensures musical artists get the royalties they are owed from streaming services, confronted him in 2023. They thought something smelled fishy, so they looked into the billions of streams associated with Smith. He denied any wrongdoing. Now he faces up to 20 years in prison for each of the charges.