Michael Smith, 54, of Cornelius, North Carolina, was sentenced to 18 months in federal prison after pleading guilty to conspiracy to commit wire fraud for manipulating music-streaming royalties across four major platforms. A federal judge also ordered him to forfeit $8,091,843.64, the agreed loss amount connected to a scheme prosecutors say defrauded Spotify, Apple Music, Amazon Music, and YouTube Music between 2017 and 2024.
Prosecutors described the case as the Justice Department’s first criminal prosecution involving AI-assisted music-streaming fraud.
How the Scheme Worked
Smith used AI tools to generate hundreds of thousands of songs, then deployed automated bot accounts to stream them billions of times, mimicking genuine listeners according to prosecutors. Distributing artificial plays across an enormous library made individual tracks appear less suspicious than concentrating activity on a small number of songs, court materials indicate.
The scale is striking: in April 2023 alone, Smith’s tracks reportedly received 80.9 million YouTube Music streams from family-plan accounts, per government sentencing materials. For context, Taylor Swift’s entire catalog received 9.3 million streams in that same category during the same period, according to the same materials. Court and government records describe bot networks ranging from roughly 1,040 to as many as 10,000 accounts across different stages or measurements of the operation.
What This Costs Legitimate Artists
The Justice Department said the scheme defrauded music platforms and musicians of royalty payments that should have flowed to artists whose work consumers actually played. Artificial streaming activity can also inflate perceived popularity and skew listener metrics, though the precise effects on recommendation systems and payment allocation go beyond what the DOJ materials alone establish.
The use of generative AI appears to have supplied the volume of content needed to make the automated streaming activity harder to detect. That combination allowed the alleged fraud to operate at a scale that would have been far more difficult through conventional music production, based on the scale of output described in court materials.
The Detection Problem
Streaming platforms must distinguish coordinated bot activity from legitimate high-volume listening without unfairly penalizing independent artists or background-music creators with unusual but authentic patterns. Platforms may need to evaluate behavioral signals such as account relationships, playback timing, and catalog-wide activity. Stream counts for individual tracks alone are insufficient to identify this kind of operation.
Legal Context
Smith pleaded guilty in March 2026 to one count of conspiracy to commit wire fraud, a charge carrying a maximum of five years, according to the Justice Department. He also received two years of supervised release. The case illustrates that current criminal law can reach AI-enabled manipulation without legislation written specifically for generative music fraud, as the New York Times reported.
A Note on the Language
The title of the official sentencing announcement from the U.S. Attorney’s Office for the Southern District of New York used the phrase “super intelligence,” while the body of the same document also uses “artificial intelligence.” A broader claim linking that phrasing to a White House policy directive is not independently verified and should be treated as unconfirmed.
What Comes Next
For legitimate artists using AI tools, the outcome underscores how that same technology can be turned against the royalty systems that sustain working musicians. Streaming services now face documented evidence that AI-assisted fraud can reach industrial scale, adding pressure to strengthen detection without creating new obstacles for creators working in good faith.
