A North Carolina man who prosecutors say used artificial intelligence to generate hundreds of thousands of songs, then deployed an army of bots to stream them billions of times, was sentenced on Tuesday to 18 months in federal prison for a royalty fraud that ran for seven years.

Michael Smith, 54, was sentenced in Federal District Court in Manhattan by Judge John G. Koeltl after pleading guilty to one count of conspiracy to commit wire fraud. Smith was also sentenced to two years of supervised release and ordered to forfeit $8,091,843.64, according to a press release issued by the U.S. Attorney’s Office for the Southern District of New York.

The Justice Department called the case the first criminal prosecution of a music streaming fraud assisted by what it described as “super intelligence” technology, a term the department used throughout its announcement without specifying which tools Smith employed.

“Michael Smith exploited super intelligence technology to generate a fraud,” Jamie McDonald, the United States attorney for the Southern District of New York, said in a statement. “By flooding music streaming platforms with automated bots in the place of consumers, and fake songs in the place of creativity, Smith robbed millions in royalty payments from genuine artists and their fans.”

An estimated $4.6 billion in annual artist revenue could be lost to AI-generated music by 2028, according to a recent economic study conducted by CISAC and PMP Strategy with participation from the French streaming service Deezer.

The case offers a glimpse of how songs made using generative AI can be turned against the economics of the modern music business, where payouts to songwriters and performers are drawn from a shared pool of money and divided in proportion to listening.

That structure is the vulnerability that prosecutors said Smith exploited. Each time a track is streamed on a service like Spotify or Apple Music, the songwriter, the performer and in some cases other rightsholders are entitled to a small payment. Because that money is pulled from a fixed pool, fake streams do not create new revenue but instead divert it from artists whose songs were played by real listeners.

According to prosecutors, Smith’s scheme ran from 2017 to 2024 and unfolded in three steps. He created thousands of accounts on the streaming services, using fake email addresses and fraudulently obtained debit cards. He then used software to make those accounts stream songs that he owned, most of which were created with AI, before collecting the royalties.

At times, prosecutors said, Smith ran as many as 10,000 of the bot accounts simultaneously. To avoid detection, he spread the automated listening across thousands of songs so that no single track would draw the kind of anomalous traffic that might alert the platforms.

That strategy required volume. In an email cited by prosecutors, Smith said he needed “a TON of songs fast” to get around the platforms’ “anti fraud policies.” By June 2019, Smith and his co-conspirators were reportedly earning roughly $110,000 every month from music royalties stemming from “hundreds of thousands” of songs.

In April 2023, the entire catalog of Taylor Swift drew 9.3 million streams on YouTube Music from family plan listeners, according to the press release. In the same month, Smith’s bot accounts streamed his AI-generated music 80.9 million times.

In all, prosecutors said, the scheme brought him more than $8 million in royalties. McDonald said his office was “committed to ensuring the integrity of all markets, and protecting the public from those who use super intelligence for fraud.”

https://edm.com/news/man-sentenced-18-months-prison-ai-music-fraud-8-million-streaming-royalties/

(http://www.autoadmit.com/thread.php?thread_id=5912525&forum_id=2#50188116)