On October 6, 2026, Judge John G. Koeltl of the federal court in Manhattan sentenced 54-year-old Michael Smith of Cornelius, North Carolina, to 18 months in prison followed by two years of supervised release. Smith, who pleaded guilty to the charges, was ordered to forfeit over $8 million in illicit royalties. This case marks the first time a federal prison sentence has been handed down in the United States for music streaming fraud, a scheme executed by siphoning small amounts of money from the shared royalty pools of platforms such as Spotify, Apple Music, Amazon Music, and YouTube Music over a seven-year period.
How did the bot network generate 661,000 streams per day?
The mechanics of the operation were detailed in an email Michael Smith sent to himself on October 20, 2017. The spreadsheet outlined a system comprising 52 cloud computing accounts, each hosting 20 virtual listener accounts for a total of 1,040 bots. Smith calculated that each bot could play approximately 636 songs daily, resulting in 661,440 streams per day. Based on a conservative estimate of half a cent per stream, he projected earnings of $3,307 per day, roughly $100,000 per month, and $1.2 million annually.
Building this infrastructure was labor-intensive. Smith purchased email accounts with fake identities and employed people domestically and abroad to complete registration forms. By May 2017, he instructed a partner to generate false names and addresses, ensuring they appeared consistent with family subscription packages, which were more cost-effective. He utilized services that provided bank cards to fictional employees to manage payments, while employing virtual private networks (VPNs) to mask the source of his traffic and avoid detection.
Why did the operation require hundreds of thousands of songs?
Michael Smith understood that streaming a single track millions of times in one night would trigger the fraud detection algorithms of streaming platforms. To bypass these security measures, he spread the bot activity across a vast catalog of tracks. By distributing the streams across thousands of songs, the number of plays per track remained low enough to avoid scrutiny, while the cumulative revenue grew.
Initially, Smith tested the method with his own limited library and that of a music promoter, but neither provided the necessary volume. Prosecutors state that he eventually purchased hundreds of thousands of AI-generated songs to sustain the scheme. The scale of the fraud was starkly illustrated in April 2023: while Taylor Swift’s entire catalog generated 9.3 million streams on YouTube Music family plans, Smith’s bots generated 80.9 million streams on the same platform using AI-produced tracks.
How does streaming fraud affect the distribution of royalty payments?
Streaming platforms do not pay a flat rate per stream. Instead, they utilize a pro-rata model where monthly subscription and advertising revenue in a given country is pooled and distributed to rights holders based on their share of total streams in that market. Because the pool size is determined by total revenue rather than total streams, every instance of bot-generated listening dilutes the earnings of genuine listeners and legitimate artists. This is how the prosecution defined the financial harm caused by Smith’s actions.
Defense attorneys argued that because the losses were distributed among millions of rights holders, the impact on any single victim was negligible. Prosecutors countered that while the losses were diffuse, the gains were concentrated in Smith's pockets, evidenced by the detailed profit-and-loss spreadsheets he maintained for each bot.
What are platforms doing to combat AI-generated streaming fraud?
Deezer was among the first platforms to identify and label music produced entirely by artificial intelligence. By June 2026, such content accounted for over 50 percent of daily uploads on their platform. Despite the volume of uploads, these tracks accounted for only 1 to 3 percent of total streams. Deezer noted that in 2025, up to 85 percent of streams for these AI tracks were identified as fraudulent.
Spotify introduced a new eligibility threshold in April 2024, requiring a track to reach at least 1,000 annual streams and a certain threshold of unique listeners to earn royalties. This policy aims to discourage schemes that upload thousands of tracks to collect small payments from each. Additionally, Spotify reported that it removed over 75 million spam tracks in the 12 months leading up to September 2025. Despite warnings dating back to 2018, Smith continued his operations until the Mechanical Licensing Collective halted his payments in March 2023, eventually leading to his arrest and conviction.
Notes and Sources
- US Department of Justice, Southern District of New York (October 6, 2026). "North Carolina Man Sentenced To 18 Months In Prison For Super Intelligence-Assisted Music Streaming Fraud."
- United States v. Michael Smith, indictment, 24 Cr. 504, US District Court for the Southern District of New York (2024).
- US Attorney’s sentencing memorandum, United States v. Michael Smith (September 2026).
- Spotify, Loud & Clear (2023). "How is streamshare calculated?"
- Bill Donahue, Billboard (October 6, 2026). "AI Music Streaming Fraudster Sentenced to 1.5 Years in Prison for $8M Royalties Scam."
- Spotify Loud & Clear Turkey report (June 2025).
- Deezer (July 21, 2026). "AI Music Tops 50% of Daily Uploads on Deezer"; Ivan Mehta, TechCrunch (July 21, 2026).
- Deezer (January 29, 2026). "Deezer confirms 85% of AI streams are fraud, now demonetized."
- Spotify for Artists. "Track monetization eligibility" (April 2024).
- Spotify Newsroom (September 25, 2025). "Spotify Strengthens AI Protections for Artists, Songwriters, and Producers."
