Man Who Used AI Songs and Bots to Outstream Taylor Swift Nearly 9 to 1 Jailed for $8M in Royalty Fraud

A North Carolina man who used AI-generated songs and thousands of automated accounts to manipulate music streams has been sentenced to 18 months in prison after fraudulently obtaining more than $8 million (£6 million) in royalties. Michael Smith was also ordered to forfeit $8.09 million (£6.12 million) and serve two years of supervised release.

The scale of the scheme was starkest in April 2023, when Smith's bot accounts generated 80.9 million YouTube Music family-plan streams for his AI-generated music, compared with 9.3 million streams for Taylor Swift's entire catalogue in the same category. That put Smith's artificial activity at about 8.7 times Swift's total.

The case highlights a developing problem for streaming platforms: AI can make it cheap to manufacture vast quantities of music, while automation can manufacture the appearance of an audience.

How Smith Built an Artificial Audience

Smith's scheme operated from 2017 to 2024 and relied on three elements: thousands of fraudulent streaming accounts, software that repeatedly played songs, and a huge catalogue of music that could absorb the resulting streams without making any single track look unusually popular.

At times, as many as 10,000 bot accounts were operating simultaneously. Smith used fake email accounts and fraudulently obtained debit cards to create them, then spread automated streams across thousands of songs to reduce the risk of detection.

AI solved another logistical problem. Smith needed hundreds of thousands of songs to keep the operation moving, and used AI to generate the necessary tracks. The resulting songs were streamed billions of times by the bot accounts.

Why Fake Streams Affect Legitimate Artists

Streaming royalties are distributed from a pool of money, meaning artificial listening can alter how that pool is allocated among rights holders. The US Justice Department says fraudulent streams divert money from musicians and songwriters whose music was genuinely played by consumers.

That makes the Smith case different from a dispute over whether AI-generated music should be allowed on streaming services. The criminal conduct centred on creating false accounts and artificial listening activity to obtain royalty payments. The distinction could become increasingly important as AI-generated music becomes more common.

AI Music Is Making Streaming Fraud Easier to Scale

IFPI's 2026 Global Music Report shows why the problem extends beyond one criminal case. Deezer reported receiving more than 60,000 fully AI-generated tracks every day in January 2026, while saying up to 85% of streams on AI-generated music on its service were fraudulent in 2025.

IFPI said generative AI has industrialised artificial content and artificial listening, making large-scale fraud cheaper and faster while potentially making it harder to detect. The organisation has called for stronger identity verification, content vetting, detection systems, and information-sharing between streaming platforms and distributors.

The Sentence Sets a Clearer Line for AI Music

Smith pleaded guilty on 19 March 2026 to one count of conspiracy to commit wire fraud before US District Judge John G. Koeltl. He was sentenced on 6 October 2026. The significance of the case is therefore less about whether machines can make music and more about whether automated systems can be used to manufacture the commercial activity surrounding it. For streaming businesses, that creates a tougher challenge: distinguishing legitimate AI-assisted creativity from artificial audiences designed to trigger royalty payments.