The AI Music 'Flood' Has Arrived: 50,000 Tracks a Day on Deezer, How Will Streaming Change?
AI is not only 'streamlining' the music industry but also shaking the foundations of the source of value (who created it/what it was created with) and revenue distribution (who gets paid). Bloomberg Tech: Europe (2026/1/9) organized these issues into the categories of 'AI artists,' 'massive influx of generative AI,' 'rights management,' and 'labeling (disclosure).'
1. 'AI Artists' Have Become Reality: Toward an Era of Indistinguishability
The program introduces the case of AI-generated R&B artist 'Xania Monet' charting on Billboard. Billboard itself emphasized the point that 'an AI artist has entered the radio charts,' symbolizing that AI has moved from an 'experiment' to a 'market presence.'
More importantly, listeners are almost unable to distinguish between them. A survey by Deezer and Ipsos reported that a very high percentage of people 'cannot tell the difference between AI songs and human songs,' indicating a strong need for transparency (labeling).
1-1. The Number of 'People Who Can Compose' Explodes
With generative AI platforms like Suno and Udio, 'decent-sounding finished songs' can be created in a short time without specialized equipment or studios. As a result, while music production is being democratized, it also invites the 'oversupply' and 'dilution of revenue' discussed later.
2. The Biggest Tinderbox is Not 'AI Tools' but the 'Massive Influx of AI Songs'
AudioShake CEO Jessica Powell breaks down the issue this way: The problem is not creators using AI as a tool, but rather '100% AI-generated' content being flooded into streaming services, stealing listeners' disposable time and the revenue pool.
Numbers supporting this concern have also emerged. Deezer has announced that it is receiving over 50,000 'fully AI-generated tracks' per day, reaching a scale of about one-third of the total.
2-1. Beyond 'AI Slop,' Streaming Fraud is Also Amplified
Because AI songs are easy to mass-produce, they are also compatible with 'fraud' that siphons royalties through bot streaming. In fact, Deezer has mentioned the possibility that many streams of AI-generated songs are fraudulent, making the legitimacy of distribution a focal point.
3. The Core of the Solution is 'Licensing' and 'Valuation': Licensing is Only the 'Entrance'
In the program, will.i.am expressed discomfort with works being used for training or consumption without the consent of rights holders, but concluded by acknowledging, 'Licensing is a solution.'
However, the fundamental question he posed was this: 'What is the value of those rights?'—Just like the 'how much per stream' issue in streaming, the AI era is reigniting the debate over value assessment.
3-1. In the UK, Protest is Visualized Through 'Silent Albums'
In the UK, regarding the institutional design that makes it easier for AI companies to use copyrighted works for training, a 'silent album' protest involving Damon Albarn and Annie Lennox became a hot topic. The point of contention is, 'Can creators really be protected by an opt-out system?'
4. The Next Battleground is 'Labeling x Distribution x New Experiences'
Comparing it to 'organic food,' will.i.am suggested that in the future, it will become standard for music to be labeled as 'human-made/AI-made.' This is not just ethics, but a business blueprint.
Labeling: Label AI songs and change how they are handled in recommendations and playlists (Deezer has already declared it will use tagging).
Distribution: Incorporate compensation for training data, revenue sharing for generated works, and anti-fraud measures into the system.
Experience: Using AI to 'deconstruct and re-edit' songs, etc., to re-monetize existing catalogs (like AudioShake's stem separation).
And major players are moving from 'confrontation' to 'integration.' The collaboration between Universal Music Group and NVIDIA is an example of using AI for discovery and production support while also advocating for rights holder compensation.