Photo Credit: Brett Jordan
Just how deep does the AI fraud problem run? A one-stop answer remains elusive, but Apple Music has provided a closer look at the issue’s scope – to the tune of approximately two billion “fraudulent streams” detected during 2025.
This telling stat emerged during a new Hollywood Reporter sit down with Apple VP of music, video, sports, and international Oliver Schusser. By the outlet’s description – meaning its summary of Schusser’s remarks – Apple Music last year “identified and demonetized as many as 2 billion fraudulent streams.”
And while those illicit plays might not have exclusively reached AI uploads, logic and plenty of evidence suggest that machine-made slop is comparatively well-suited for gaming the system.
Already ahead of the curve when it comes to spotting AI audio, Deezer recently revealed that it’d flagged and demonetized roughly 85% of the relevant works’ streams due to fraud, for example.
This leads to another interesting sub-topic: On-demand streaming platforms’ varying approaches to tackling fraud.
As reiterated by Schusser, Apple Music, besides demonetizing fake streams, has since 2022 been slapping “fraudsters” with a sliding fine of between 5% and 25% of would-be royalties.
Now, as of yesterday, these percentages have doubled to 10% and 50%, per Schusser, who attributed the move’s timing in part to the ongoing AI audio avalanche. In general, tackling fraud is definitely a positive – though the retooled penalty policy raises questions about the underlying approach.
Among different things, if streaming fraud is egregious enough to warrant a 50% penalty, might it be time to ban the involved “artist(s)” and uploader altogether?
As tracked by DMN Pro, on Spotify and rival DSPs, AI slop peddlers have weaved elaborate featured-artist and playlist webs to drive streams. If properly handled, one instance of verified fraud could probably enable a number of justified takedowns.
An adjacent consideration: Also as covered by DMN Pro, judging by their upload volume and per-work streams, crafty AI slop specialists are avoiding directing thousands upon thousands of plays to individual tracks.
In other words, depending on how it’s applied, the adjusted penalty might only result in modest fines; then there’s the fact that AI “artist” profiles are seemingly endless and therefore essentially disposable.
DMN contacted Apple Music for additional enforcement details – including about when exactly it’s time to upgrade a fine to a straight ban – but didn’t immediately receive a response.
Elsewhere in the interview, Schusser took a shot at ad-supported listening, called out competitors that are “really struggling” with fraud, and emphasized the perceived “lot of work” needed in the industry to arrive at a universally accepted definition of AI music.
