Apple Music raises penalties on fraudulent streams
Since 2022, Apple Music has used a sliding penalty system. It deducted between 5% and 25% of potential royalties from accounts linked to fraud. This week, however, the company doubled down. Penalties now range from 10% to 50% of would-be royalties. According to Apple, the explosive surge in AI-generated content drove the stricter rules. In short, the old deterrents no longer stopped bad actors who flood services with synthetic music.
Data from Deezer shows how big the wave is. The French service now receives about 60,000 AI-generated tracks a day, or 39% of all uploads. Moreover, Deezer flagged up to 85% of streams on fully AI-produced music as fraudulent in 2025. By contrast, the fraud rate across its whole catalog was just 8%. Clearly, AI music has become the vehicle of choice for organized fraud.
How modern streaming fraud works
In the past, fraudsters uploaded a few tracks and used bots to stream them millions of times. Those spikes were easy to detect. Today, the methods are smarter. Bad actors use generative tools like Suno and Udio to create thousands of unique tracks at almost no cost. Then they spread fake plays across the whole catalog. Each song earns only a few thousand plays, so it stays under detection limits. Yet together, these libraries can siphon millions of dollars a year.
The IFPI valued the global recorded music market at $20.4 billion in 2024. Conservative estimates put fraud at 1% to 3% of all streams. Some specialists, however, argue it approaches 10% on certain platforms. At those levels, fraud could divert anywhere from $200 million to over $2 billion a year from real artists and rights holders.
Apple’s two billion blocked streams equal roughly $17 million in royalties. Schusser framed this as simple redistribution: money taken from fraudsters goes back to honest creators.
A low-cost crime with real victims
AI has lowered the barrier to entry. Making synthetic tracks needs no training, gear or studio. A single person with AI tools and basic scripts can produce thousands of songs a day. As a result, both organized crime and lone opportunists find the scheme attractive.
Detection has turned into an arms race. Apple monitors streaming patterns, play locations, listening time, skip rates and many other signals. It also shares intelligence with distributors and chart providers. Meanwhile, fraudsters adapt with residential proxies, human-like playback and many fake artist profiles.
The human cost is real. Independent artists see their small streaming income diluted. Emerging musicians also struggle for visibility when fake tracks crowd recommendation systems.
One case made headlines. In September 2024, federal authorities charged North Carolina resident Michael Smith with wire fraud, conspiracy and money laundering. Prosecutors say his seven-year scheme earned over $10 million. He allegedly used hundreds of thousands of AI tracks under fake names like “Calm Knuckles” and “Zygophyceae.” According to the indictment, over a thousand bot accounts streamed them, earning about $3,300 a day at the peak.
How other platforms respond
Spotify, with over 500 million users, says less than 1% of its streams are fraudulent. Even so, it reportedly removed 75 million “spammy” tracks in September 2025. Critics note that a small percentage still means huge absolute numbers.
Deezer has taken the toughest line. It licensed its AI detection technology to SACEM, France’s copyright society. It also keeps fully AI-generated music out of algorithmic recommendations and editorial playlists.
These different approaches reflect a wider debate. Some voices want to ban AI uploads entirely. Others warn that bans would hurt artists who use AI creatively. Some experts also propose deeper fixes, such as minimum stream thresholds, user-centric payments or artist verification.
What comes next
Both sides will keep escalating. AI music will sound even more human, and bots will mimic real listeners more closely. Therefore, detection must move toward deeper behavioral analysis. The IFPI has called for global cooperation among platforms, distributors, tech companies and law enforcement.
For working musicians, the crisis adds yet more pressure. Per-stream rates are already tiny, and fraud makes them smaller. Consequently, many creators lean on live shows, merch, fan platforms and sync deals. Ultimately, Apple Music’s disclosure proves that detection works. At the same time, it confirms how massive the theft remains.