The US Copyright Office Just Declared War on Fake Streams

The two announcements landed 24 hours apart, and together they mark a turning point for the music business. On Monday, October 6, a federal judge in Manhattan sentenced a North Carolina musician to 18 months in prison for flooding streaming platforms with AI-generated songs and bot accounts. On Wednesday, October 7, the US Copyright Office opened a formal public inquiry into music streaming fraud, asking Congress-sized questions about who is stealing from whom in the streaming economy.

The timing is not a coincidence. The notice of inquiry, Docket No. 2026-6, names the case as its headline example. But the inquiry is not about one man. It is the first time the federal government has asked, systematically and on the record, whether the platforms and distributors that run the global streaming economy are doing enough to keep it honest.

What Washington is asking

The inquiry traces back to a four-page letter sent on May 21 by Representative Scott Fitzgerald, a Wisconsin Republican on the House Judiciary Committee. Fitzgerald asked the Copyright Office to examine, in his words, the contributory effects of artificial intelligence in music streaming fraud. He was careful not to blame the technology itself: there is, he wrote, nothing inherently wrong with using AI as a creative tool in music development. The problem, he argued, is that generative AI serves as a catalyst for criminals to perpetrate streaming fraud at a scale no human could match.

The notice poses ten sets of questions. Some ask for facts: how fraud schemes work, how common bot farms, click farms, playlist manipulation, and fraudulent marketing companies are, and what the damage costs in dollars. Others are more pointed. One invites the industry to say which of its own policies or practices may be inadvertently contributing to the problem. Another asks which anti-fraud measures actually work, from trusted distributors and customer checks to royalty clawbacks, fines, takedowns, and outside vendors such as Beatdapp. A third floats the idea of a shared database to track known instances of streaming fraud across the industry.

The final questions are the ones that matter most. The Office wants to know whether voluntary fixes are enough, or whether Congress needs to legislate. The notice footnotes three bills as reference points: the Block Bad Electronic Art and Recording Distributors Act discussion draft from 2025, the American Copyright Protection Act of 2026, and the DEFEND IP Act.

When AI-generated outputs compete against human-made works, they divert money from a finite royalty pool that would otherwise be distributed to actual rights holders.

How fake streams take money from real artists

The mechanics of the theft are what make this story a policy story rather than a simple crime story. Streaming services set aside a percentage of their revenue as royalty pools, and under the standard pro-rata system, the more a song is played, the more its rights holders receive. That means the pool is effectively fixed each month: every bot play takes a slice, and every real artist share gets a little smaller. The Copyright Office calls this the dilution of the royalty revenue pool, and it lists the dilution as the primary harm of streaming fraud.

The harm does not stop at the check. The inquiry notice also lists secondary injuries. Fake plays push real songs down playlists and algorithmic recommendation queues, which determines what new listeners discover. And the distorted streaming data then feeds the business decisions artists and venues make about where to tour, what to merchandise, and which markets to invest in. A fake stream is never just a number; it is a signal that poisons the information the whole live music economy runs on.

The numbers that forced the issue

Fitzgerald evidence is mostly supplied by Deezer, the French streaming service that has been the most aggressive public reporter of fraud data. Its January report found that up to 85 percent of streams on AI-generated tracks were fraudulent in 2025. By April, nearly 44 percent of its daily uploads were AI-generated. Across its entire catalog, Deezer says fraud made up 8 percent of all streams last year. The letter adds that Apple Music reportedly demonetized 2 billion fraudulent streams in 2025 alone.

Then there is the case that made the problem impossible to ignore. Michael Smith, 54, of Cornelius, North Carolina, built a scheme that ran from 2017 to 2024: hundreds of thousands of AI-generated songs, as many as 10,000 automated bot accounts, and billions of fake plays. The Department of Justice says the operation generated more than 8 million dollars in royalties that would otherwise have remained available to legitimate artists. Smith pleaded guilty in March to conspiracy to commit wire fraud; the court ordered forfeiture of 8,091,843.64 dollars on top of the prison term.

The industry is already scrambling

The inquiry arrives while the industry is already experimenting with fixes, and the notice asks which ones work. Apple Music doubled its fraud fees in February 2026 to a sliding scale of 10 to 50 percent, according to a report the notice cites. Spotify charges distributors 10 dollars per track where fraudulent streams are detected. The IFPI launched its Streaming Integrity Initiative in September. ORCA, an indie label think tank, is pushing for a per-track upload fee that would make mass uploading economically painful. And LANDR, the distribution platform, says Spotify-flagged fake streams on its releases fell 95 percent after it tightened its own checks.

The notice also gestures at where cooperation breaks down. One question footnotes the lawsuit Universal filed against DistroKid as the counterexample to labels and distributors working together. And across the Atlantic, Italy competition authority opened an investigation on October 7 into the terms of service of Suno, the AI music generator: a consumer contract case about unilateral term changes and copyright language that may ask users to waive moral rights. The regulatory pressure on AI music is now coming from both sides of the Atlantic at once.

What happens next

The inquiry is a comment process, not a court case, and anyone can participate. Initial comments are due November 23, 2026, at 11:59 pm Eastern, filed through regulations.gov; reply comments are due December 21. Short comments can be typed directly into the form, up to 5,000 characters, and longer filings can be uploaded as a single file. Everything submitted becomes public, names included.

There is precedent for a Fitzgerald letter turning into real policy work. His September 2024 letter, co-signed with Jim Jordan and Darrell Issa, led the Office to open a 2025 inquiry into performance rights organizations that drew more than 5,000 comments, and the Office sent Congress a summary in November 2025. This time the participants are likely to include labels, streaming services, distributors, and independent artists, all arguing over a simple question: is voluntary monitoring enough, or does the streaming economy need federal rules?

For independent artists, the stakes are unusually concrete. If they do not file, Congress will hear only from the companies that run the pool. The fraud economy grew because the economics rewarded it: infinite tracks, infinite plays, and a fixed pot of money. The inquiry is the first serious attempt to change those economics on the record. The deadline is November 23.