North Carolina Man Sentenced to 18 Months for AI Song Streaming Fraud, $8 Million Seized
On October 6, the U.S. Attorney's Office for the Southern District of New York announced that Michael Smith (54) from Cornelius, North Carolina, was sentenced to 18 months in prison, two years of supervised release, and ordered to forfeit $8,091,843.64 for conspiracy to commit wire fraud.
Smith pleaded guilty on March 19, 2026. The case revealed that from 2017 to 2024, he used up to 10,000 bot accounts on Spotify, Apple Music, Amazon Music, and YouTube Music to generate billions of false plays for his AI-generated songs, earning over $8 million in royalties.
He created accounts using fraudulent debit cards and fake emails, generating hundreds of thousands of songs with AI, and dispersed plays to avoid detection of abnormal activity on individual tracks. In April 2023, his bot accounts generated 80.9 million plays on YouTube Music, while Taylor Swift's entire catalog received only 9.3 million plays during the same period.
This case is referred to by the Justice Department as the first criminal case of AI-assisted music streaming fraud in the U.S. Prosecutors had recommended at least 46 months in prison, while the probation office suggested 24 months, ultimately resulting in an 18-month sentence. Smith's lawyer argued for a not guilty plea, claiming minimal harm to individual artists.
The mechanical licensing collective flagged his data anomalies and withheld related royalties. Smith had stated in emails that he needed a "large number of songs" to evade platform detection and referred to AI-generated tracks as "instant music."
Market mechanisms drive the event through the distribution of streaming royalties based on play counts, with false plays directly diluting the shares of real rights holders. The beneficiaries are those who exploit royalties through bots and AI, while the burden falls on real artists, songwriters, and the platform's anti-fraud systems sharing the royalty pool.
Source: Public Information
ABAB AI Insight
Since 2017, Smith initially inflated plays for his own music, later collaborating with AI music providers to obtain thousands of generated songs weekly, shifting from concentrated plays of singles to dispersed plays across large catalogs to evade platform detection. After being charged in 2024, he pleaded guilty in 2026, becoming the first in the U.S. to receive a criminal conviction for AI streaming fraud.
In terms of capital pathways, he operated bots in bulk through cloud service accounts, maintaining low costs for family plans with fraudulent debit cards, converting AI-generated songs with zero marginal costs into a continuous cash flow of royalties. Once royalties were paid from the platform to his personal account, they were deemed illegal proceeds to be forfeited by the court.
Comparable cases include a Danish individual sentenced to 18 months (later increased) in 2024 for using bots to inflate 689 tracks, as well as the traditional music industry's long-standing crackdown on "play farms." Currently, streaming is at a stage where "royalty pools are distributed based on play counts, coinciding with an explosion of AI-generated content," making detection more challenging due to the increased scale of catalogs.
Structurally, this falls under regulatory changes and shifts in pricing power. AI transforms music production from scarce creation to infinitely replicable inputs, while bots shift consumption signals from real users to programmable traffic. The mechanism is that royalties are pooled and distributed based on total play counts, and any scalable false plays will directly squeeze the earnings of real rights holders, making platform anti-fraud measures and judicial prosecution pathways to recalibrate distribution rules.
ABAB News · Cognitive Laws
Zero-cost content plus programmable traffic equals royalty arbitrage
Dispersed plays are the basic strategy to evade detection
Real creators bear the brunt when the royalty pool is diluted