Artificial intelligence is no longer limited to writing text, generating images, or producing polished video in seconds. It is now moving decisively into the music industry, and not in a marginal way. Streaming platform Deezer revealed that roughly 44% of all new songs it receives each day are already generated by AI, a figure that dramatically changes the scale of the debate around creativity, automation, and authenticity in digital music.

According to the company, that amounts to nearly 75,000 AI-generated songs uploaded to its platform every day, or more than 2 million tracks per month. The figure alone is striking. But what is even more unsettling is not merely that platforms are being flooded with synthetic music, but why so much of that content is being uploaded in the first place. Deezer says the main use of this wave of AI-generated songs does not appear to be artistic or experimental, but fraudulent.

According to Deezer, while AI-generated music now accounts for nearly half of new daily uploads, its actual consumption remains relatively low, between 1% and 3% of total streams. However, the company also says that roughly 85% of those AI-related streams are detected as fraudulent and are demonetized. That detail changes the meaning of the trend. This is no longer only about artists competing with algorithms or listeners discovering songs created by machines. It is also about a possible abuse ecosystem where thousands of songs are uploaded to exploit recommendation systems, artificially inflate plays, or siphon money from royalty pools.

That point matters because for a long time, the public conversation around music and AI has focused on creative and ethical questions: whether a song made by AI can be considered art, whether it displaces human musicians, or whether copyrighted works should be used to train models. All of that still matters. But what Deezer is showing is that the challenge is no longer just philosophical. It is also about platform integrity, defense against automated spam, and economic sustainability for human creators.

The company has tried to position itself as one of the most aggressive players in this response. Deezer says it is the only streaming platform openly labeling AI-generated music transparently for users. It also says those songs do not appear in algorithmic recommendations or editorial playlists. The company further announced that it will stop storing high-resolution versions of such content, a move that suggests it is trying to reduce both infrastructure costs and incentives for opportunistic uploaders.

The growth has also been extremely fast. Deezer says that in January 2025 it was detecting around 10,000 AI-generated songs per day. Now the number is around 75,000. That leap in just over a year makes it clear that synthetic music is no longer a curiosity but a structural pressure on streaming platforms. Tools such as Suno, Udio, Lyria, and other generative systems have drastically lowered the barrier to producing full songs. What once required musicians, studio time, mixing, and labor can now be produced in minutes with a strong enough prompt.

That does not mean all AI-generated music is worthless or that all AI-assisted creation is illegitimate. There are clearly legitimate and creative uses for these tools, from music prototyping to support for independent creators. The problem appears when scale, opacity, and automation allow the system to be exploited. If a platform receives tens of thousands of synthetic songs per day, the challenge is no longer simply curating interesting content. It becomes about protecting the incentive structure so the platform does not turn into a factory of noise, fraud, and saturation.

There is also another important layer: the relationship between visibility and trust. If users feel that platforms are being flooded with automatically generated songs designed to manipulate metrics, the listening experience changes. Recommendations lose value, human artists compete with industrialized content, and the line between authentic discovery and algorithmic manipulation becomes increasingly blurred. In that sense, AI-generated music does not only challenge musicians. It also challenges the credibility of platforms that claim to help users discover meaningful work.

TechCrunch and Ars Technica emphasized that exact point: even if actual listening share remains low for now, the growth in upload volume and the percentage of suspected fraud show that this is not a minor issue. It is an early warning of how AI can transform not only cultural production, but also economic systems that depend on digital signals, automated monetization, and algorithmic visibility.

The conclusion is uncomfortable but necessary. The biggest problem with AI-generated music may not simply be that a machine can compose convincing songs. The real problem may be that this capability, at massive scale, turns music platforms into fertile ground for spam, manipulation, and the dilution of human creative value. And if that already represents 44% of new uploads on a major platform, the industry is not facing a future possibility. It is facing a transformation that has already begun.