Artificial Intelligence

28% of music on Deezer is produced by Ai: the end of art or a new era?

Deezer and others seek to regulate the use of AI to protect artists and revenues in the future of music

28 September 2025

2' min read

Major music streaming platforms are facing a new and controversial phenomenon: the massive presence of AI-generated tracks in their catalogues. This has triggered a wave of concern in the global music scene, which perceives AI-generated music as a threat. But is this really an inevitable risk for the future of music?

The meteoric rise of AI music

The music streaming platform Deezer has sounded a clear alarm, revealing significant data: every day it receives more than 30,000 tracks entirely generated by artificial intelligence. This number is rising sharply, from 10,000 tracks per day in January to over 30,000 at present.

According to the platform, up to 70 per cent of the listens to these tracks were fraudulent, aimed at illicit earnings through royalty payments. Although this music accounts for only 0.5 per cent of total listens, its purpose does not appear to be artistic, but purely economic. Alexis Lanternier, CEO of Deezer, said the company wants to be a leader in dealing with the problem to minimise any negative impact on artists and fans.

Industry alarm: the 'AI Slop' phenomenon

This phenomenon has been christened 'AI Slop': it refers to low-quality content, such as music, images or videos, generated en masse by artificial intelligence without any real creative purpose, but only to produce huge amounts of files effortlessly.

This amount of content is flooding several platforms, including music streaming platforms. For example, 'Aventhis', a dark country musician entirely generated by AI, saw his ratings halve after his nature was revealed, demonstrating public fear and rejection.

The challenges for the future of music

Deezer has launched a tagging system to clearly show users the origin of AI-generated tracks and exclude them from editorial playlists. However, alarm remains high. According to a study by CISAC and PMP Strategy, 25 per cent of creators' revenues could be at risk by 2028, with a potential loss of EUR 4 billion.

It should also be emphasised that the earnings generated by these tracks are doubly fraudulent, as the generative technology was trained using the music of existing artists, without them being paid for the use of their data.

What to expect?

The discourse I always make is that of a conscious and ethical use of artificial intelligence, giving account and credit if an artist has unknowingly fed a music prompt site. On the other hand, I hope that artists, rightly irritated by the attitude of those behind these tracks, will be able to use AI as a tool to renew their production and also be able to 'play' with the present.