Taylor Swift has millions of devoted fans around the world. An American musician named Michael Smith discovered that there was another way to achieve extraordinary streaming numbers: create the music with artificial intelligence, invent the audience and leave the listening to computers.
Between 2017 and 2024, Smith built an operation involving hundreds of thousands of songs and thousands of automated accounts on platforms including Spotify, Apple Music, Amazon Music and YouTube Music. The songs were generated using AI, while software bots played them continuously, creating the appearance of an enormous and enthusiastic audience.
At one point, the results were remarkable. In April 2023, Smith's operation generated approximately 80.9 million YouTube Music family-plan streams, compared with 9.3 million for Taylor Swift's catalogue in the same category.
Of course, Swift's fans were real. Smith's listeners were largely computer programmes.
The operation reportedly involved as many as 10,000 active bot accounts, with streams spread across thousands of tracks to avoid attracting attention. Instead of producing one viral hit, Smith created an entire artificial music industry, complete with songs, listeners and royalty payments.
And the financial results were anything but imaginary. The scheme generated more than $8 million in fraudulent royalties over several years.
The extraordinary numbers eventually attracted the attention of US authorities. Smith pleaded guilty to conspiracy to commit wire fraud and was sentenced in October 2026 to 18 months in prison.
Beyond the fraud itself, the case illustrates an increasingly strange reality of the digital music business. Artificial intelligence can compose songs, automated accounts can play them, and algorithms can turn those plays into apparently impressive popularity statistics.
For decades, musicians have struggled to find an audience. Smith found a way to manufacture one.
The only thing missing was someone who actually wanted to hear the music.
