- Michael Smith was sentenced to 18 months in prison for using bots to manipulate music streams and fraudulently obtain more than $8 million in royalties.
- His operation involved thousands of automated accounts that generated billions of artificial plays across music streaming platforms.
- Google says its SynthID technology has watermarked 240,000 years’ worth of AI audio and around 180 billion AI generated images and videos.
- AI music creation is not illegal by itself, but using bots to manipulate streaming figures and obtain royalties fraudulently can lead to criminal prosecution.
The rapid growth of AI generated music has created new opportunities for musicians and producers, but it has also opened the door to large scale streaming fraud. A US man has been sentenced to 18 months in prison after using thousands of automated bots to manipulate music streaming figures and collect more than $8 million in royalties.
Michael Smith, from Cornelius, North Carolina, admitted to conspiring to commit wire fraud after running the scheme between 2017 and 2024. During that period, he uploaded hundreds of thousands of AI-generated tracks to major streaming platforms, including Spotify, Apple Music and Amazon Music. He then used automated accounts to generate billions of artificial plays, making his music appear far more popular than it really was.
The case highlights a growing challenge for the music industry. AI tools can now produce large volumes of music quickly, while streaming platforms rely on listening activity to distribute royalty payments. When automated systems are used to manufacture that activity, legitimate musicians and songwriters can lose out.
Smith’s sentence is also a reminder that creating music with AI is not, by itself, the offence in this case. The criminal conduct centred on manipulating streaming numbers and obtaining money through fraudulent activity.
According to the US Department of Justice, Smith was ordered to forfeit $8,091,843.64. He will also serve two years of supervised release after completing his prison sentence. Although he faced a potential sentence of up to five years, the court ultimately imposed an 18 month term.
Thousands of Bots Helped Generate Billions of Plays
Smith’s operation depended on a combination of automated streaming accounts, cloud services and VPN connections. These tools helped him distribute artificial listening activity across platforms, making it harder to identify the plays as part of a single coordinated operation.
According to details reported about the case, Smith used 52 cloud service accounts, each providing access to 20 bots. That gave him control of 1,040 bots under the arrangement described in those reports.
The automated accounts were reportedly capable of streaming around 636 songs each day. Across the network, that would amount to approximately 661,440 plays daily, creating an enormous volume of activity without requiring real listeners.
The Justice Department cited an even larger operation, stating that Smith deployed as many as 10,000 bots at a time.
One example illustrates the scale of the manipulation. In April 2023, Taylor Swift’s music received 9.3 million plays on YouTube Music, according to the figures cited in the case. Smith’s tracks reportedly generated 80.9 million plays on the same service during that month.
Those numbers underline why artificial streaming has become a serious concern for music services. Royalty systems are designed to reward music that people actually listen to. When bots inflate the figures, the resulting payments can be diverted away from artists who earn their audiences legitimately.
Smith was initially charged in September 2024 with wire fraud conspiracy, wire fraud and money laundering conspiracy. In March 2026, he pleaded guilty to a single count of conspiring to commit wire fraud.
The case demonstrates that streaming fraud can carry serious criminal consequences when automated activity is used to deceive platforms and obtain royalty payments.
Google’s SynthID Expands AI Audio Detection
The fraud case comes as Google continues to expand its efforts to identify AI generated content. The company has revealed that its SynthID technology has watermarked audio equivalent to around 240,000 years of continuous listening, alongside approximately 180 billion AI generated images and videos.
SynthID embeds identifying information into AI generated content. The watermark is designed to remain detectable even when the content undergoes certain changes, helping identify material produced using supported AI systems.
Google has also expanded access to its SynthID Detector, a tool intended to help users determine whether content contains a SynthID watermark. The tool was initially available to professional users, but Google has since broadened access.
The technology could become increasingly useful as AI generated music spreads across digital platforms. Streaming services receive enormous volumes of new material, and manually checking every upload would be difficult and expensive. Watermarking offers another way to identify AI generated content and support moderation efforts.
However, there is an important distinction between identifying AI generated material and detecting fraud. An AI watermark can help establish whether supported technology was used to create a piece of content, but it does not automatically prove that a stream is fake or that a creator has committed an offence.
Google’s announcement also does not mean that every AI generated track across every platform can be identified with certainty. Detection depends on the watermark being present and the technology being able to recognise it.
That distinction matters for music services trying to address two separate issues: the growing volume of AI generated uploads and the deliberate manipulation of streaming figures.
AI Music Is Growing, but Real Engagement Remains a Challenge
AI music generation has made it easier than ever to create tracks and distribute them online. But producing music quickly does not guarantee an audience, and the majority of uploads may struggle to attract meaningful listening figures.
Apple Music executive Oliver Schusser has previously said that more than a third of the tracks delivered to the platform were entirely AI generated, while accounting for less than 0.5% of usage. The figures suggest that a large share of AI music receives relatively little listener attention.
This creates a difficult balancing act for streaming companies. They need to accommodate new creative tools without allowing automated uploads and artificial engagement to undermine their royalty systems.
The problem is not limited to AI generated music. Streaming fraud can also involve recordings made by human artists, with bots used to inflate their apparent popularity. AI simply makes it easier to produce vast catalogues of tracks that can be uploaded in large quantities.
For legitimate musicians, the consequences can extend beyond lost royalties. Artificial streams can distort popularity metrics, influence recommendations and make it harder for platforms to distinguish genuine audience demand from manipulated activity.
Watermarking, improved fraud detection and closer monitoring of unusual listening patterns could all help address these problems. No single measure, however, is likely to eliminate streaming fraud entirely.
Smith’s case illustrates the financial incentives behind the practice and the potential consequences when those incentives lead to criminal activity.
The central issue is not whether AI should be allowed to make music. It is whether platforms can reliably identify fraudulent behaviour in an industry where creating and distributing content is becoming faster and cheaper.
Follow TechBSB For More Updates
