According to Music Business Worldwide, the US Copyright Office opened a public inquiry into music streaming fraud, with AI generated tracks and royalty diversion squarely in scope. On October 7, 2026, The Verge reports that Michael Smith received a prison term of 18 months for a bot streaming scheme involving hundreds of thousands of AI generated songs and more than $8 million in royalties. The practical issue is whether platforms can distinguish synthetic volume from legitimate listening.
Streaming fraud enters policy and enforcement
The US Copyright Office, Music Business Worldwide reports, opened a public inquiry that makes AI’s role in streaming fraud a formal policy question. Music Business Worldwide says the inquiry seeks public comments on fraud’s effects on licensing, royalty distribution, and the economic position of human creators. The source frames the scope beyond track generation, covering systems that count usage and pay rightsholders. In the criminal case, The Verge details that Michael Smith had pleaded guilty earlier in 2026 before receiving a prison term of 18 months. The practical effect is a concrete enforcement outcome beside an open policy process, with technical checks and industry obligations still undecided.
Italy examines Suno’s contract language
According to Music Ally, the Italian Competition Authority, known as AGCM, opened an investigation into Suno’s terms of service on October 7, 2026. As Music Ally details, the authority is examining clauses that may let Suno change contract terms unilaterally, alongside copyright language that may require consumers to waive moral rights. Music Ally also says AGCM will seek input from trade associations and consumer groups through a public consultation. The practical distinction is that this is a consumer contract case, not a decision on model training. What is not yet clear is whether Suno will revise the language before the process concludes.
AudioShake prepares mixed recordings for AI training
AudioShake introduced The Refinery on October 8, 2026, a tool that AudioShake says turns finished recordings into structured data for AI training. AudioShake says the tool can separate overlapping voices and isolate dialogue, music, and background sound from a mixed recording. It is designed to recover usable training material from audio that would otherwise be difficult for machine learning systems to process, as AudioShake explains. The practical value is cleaner data, but that does not answer whether a recording was authorized for this specific reuse. What remains unresolved in the announcement is how customers will document consent, ownership, and permitted uses for recordings converted into training data.
