Copyright Office opens inquiry into AI and music-streaming fraud

The U.S. Copyright Office has opened a public inquiry into music-streaming fraud, seeking evidence on artificial plays, the economic effects on the music business and whether voluntary action, regulation or legislation could curb the practice. A notice dated Oct. 5 and published in the Federal Register on Oct. 7 answered a request from Representative Scott Fitzgerald to examine fraud across digital platforms and its effects on royalties for sound recordings and musical works. Initial comments are due by 11:59 p.m. Eastern Time on Nov. 23, with reply comments due by 11:59 p.m. on Dec. 21. The inquiry arrived one day after Michael Smith was sentenced over an AI-assisted streaming-fraud operation, as this project reported earlier this month. The Copyright Office notice cites the Justice Department’s case against Smith, alleging that he created bot accounts to stream uploaded songs and collected roughly $10 million in royalties fraudulently. The office defines streaming fraud as using deceptive, inappropriate or unpermitted means to inflate a song’s or album’s play count, whether to influence public perception, obtain financial rewards or gain another improper benefit. Its questions specifically cover bot and click farms, playlist stuffing or manipulation, account hijacking and other techniques used to generate artificial streams. It also asks how fraud relates to AI-generated music, including the consequences for licensing and royalty distribution. The consultation will consider the adequacy of industry detection and enforcement efforts, as well as possible criminal-enforcement and legislative responses. The stakes extend beyond individual fraudulent payments because streaming services commonly put a portion of revenue into royalty pools. Under the prevailing pro-rata model, more plays generally mean a larger payout to a work’s rights holders. Artificial plays therefore dilute payments available for legitimately streamed music, the Copyright Office said. The agency said fraudulent activity can also influence playlists and algorithmic queues, compromise data used by artists and venues, limit discovery for lesser-known performers, and create financial or reputational harm. Its notice places domestic streaming-specific revenue for sound-recording copyright owners at about $9.5 billion, and global revenue at about $22 billion. Platforms have developed several ways to identify suspicious activity, including reviewing listening time and user geography, comparing artist streaming patterns, and measuring saves and follows against total plays. Providers have fined uploaders linked to artificial streams, while distributors have expanded identity checks, trust-and-safety teams and know-your-customer procedures. Yet the notice says fraudsters can move to another distributor after being detected. It cites industry coordination through the Music Fights Fraud Alliance, launched in 2023, and IFPI’s Streaming Integrity Initiative, whose participants committed to baseline anti-fraud practices. The review also raises whether royalty-system changes could reduce the dilution of legitimate plays. The office is considering artist-centric and user-centric alternatives to the pro-rata model as potential approaches. AI has increased both the scale and sophistication of streaming fraud, according to WIPO Magazine. Earlier schemes could rely on repeatedly playing a limited group of tracks, whereas AI-enabled operations can spread relatively modest play counts over millions of synthetic songs. WIPO also reported that AI can be used both to generate audio and to operate streaming bots or spoof digital identities.