In what prosecutors describe as the first criminal case over AI-assisted music streaming fraud, North Carolina producer Michael Smith has been sentenced to 18 months in federal prison for using thousands of bot accounts to stream his own AI-generated tracks billions of times and siphon more than $8 million in royalties from major platforms and real artists. Smith, 54, will also serve two years of supervised release and has been ordered to forfeit more than $8,091,843 in illicit earnings, marking a milestone moment in the battle against automated “stream farms” and synthetic music flooding services like Spotify, Apple Music, Amazon Music, and YouTube Music.
According to court documents from the Southern District of New York, Smith’s scheme ran from 2017 through 2024 and followed a simple but devastating playbook: create fake accounts at scale, generate music with AI tools, and unleash automated software to stream those tracks around the clock. Prosecutors say he used email addresses purchased online to spin up thousands of accounts—later dubbed “Bot Accounts”—across the major streaming platforms, then relied on what filings describe as “super intelligence” to generate hundreds of thousands of tracks credited to invented artists that he controlled. At peak, as many as 10,000 bots were running simultaneously, hammering his catalog billions of times and quietly diverting royalty money away from real musicians whose work shares the same payout pool.
The scale of the operation only becomes clear when you look at the numbers the government put on the record. In one example highlighted in the sentencing materials, Smith’s bot-driven catalog racked up 80.9 million streams on YouTube Music family plans in April 2023 alone, compared with 9.3 million family-plan streams for Taylor Swift’s entire catalog on the same service during that period—a nearly ninefold gap created by scripts rather than fans. Investigators say this kind of manipulation exploited family subscription plans, which bundle multiple listeners under one account and can make abnormal activity look more like heavy household use than industrial-scale fraud. Across the life of the scheme, Smith’s AI-generated songs were streamed billions of times, almost none of them representing genuine human listening.
Smith was first charged in September 2024 on a three-count indictment that included wire fraud conspiracy, wire fraud, and money laundering conspiracy, each carrying a potential 20-year maximum sentence. In March 2026, he entered a guilty plea to a single count of conspiracy to commit wire fraud, a charge with a five-year statutory maximum, with prosecutors emphasizing that he had “stolen” royalty payments from hardworking songwriters and performers who rely on streaming income. U.S. District Judge John G. Koeltl ultimately imposed an 18-month term—shorter than the nearly four years prosecutors argued for but still significant for a first-of-its-kind AI streaming case—along with forfeiture of the fraud proceeds and the two-year supervised release tail.
For the music industry, the case lands right in the middle of an anxiety spike over both streaming fraud and the rise of generative AI flooding services with low-effort tracks. Trade outlets and legal analysts have noted that streaming manipulation has typically involved human-made songs boosted via playlist scams or paid stream farms, but Smith’s operation shows how generative tools now let a single person fabricate enormous catalogs of disposable content, then weaponize bots to turn that junk into real money. Commentators point out that because streaming royalties are pooled and then divided among all plays, this kind of artificial inflation doesn’t just hurt platforms—it directly dilutes payments for legitimate artists by crowding their music out of the revenue share.
Tech-wise, what Smith did wasn’t science fiction; it was a dark-side remix of everyday automation. Prosecutors say he used off-the-shelf software to control swarms of account logins and loop his own tracks, while leaning on accessible AI music tools to mass-produce songs fast enough to avoid simple “repeat track” detection. The scheme underscores how subscription models and recommendation algorithms can be gamed when platforms are slow to distinguish genuine listeners from scripted traffic or to flag catalogs made almost entirely of machine-generated audio. It is the kind of exploit that sits right at the intersection of geek tech—scripts, bots, generative models—and creative industries that never expected to fight botnets in the charts.
The fallout is already prompting calls for platforms to tighten their defenses against automated listening and to rethink how they treat AI-generated music in royalty calculations. Legal observers say Smith’s conviction sends a clear signal that “super intelligence” is no defense when it is used as a tool for fraud, and that future schemes leveraging bots and generative media are likely to meet the same fate in court. For artists experimenting with AI in good faith, the case draws a new line in the sand: using algorithms as part of a creative workflow is one thing, but fabricating audiences and hijacking royalty pools is now firmly established as criminal territory.
