AI music tools explode, but mainly on the creation side
Behind the wave of AI generated music is a fast-growing industry of creation tools. According to the IMS Electronic Music Business Report 2025/26, generative AI and stem separation platforms reached USD 333 million (approx. RM1.6 billion) in revenue in 2025, with a 651% surge between 2023 and 2025 and 63 million monthly active users. These tools sit at the heart of a production shift: instead of relying only on traditional plug‑ins and studio software, creators now use AI to pull stems from existing tracks or to generate new beats, melodies and even full songs from text prompts. The report stresses that AI’s biggest impact is currently on the creation side, not on how people listen. In other words, more Malaysians may be experimenting with AI platforms to make tracks than actively seeking out AI songs on Spotify or other streaming apps.
Deezer’s AI uploads boom — but streams stay tiny
Deezer’s latest music streaming stats show just how sharply AI uploads have spiked. The company now sees nearly 75,000 AI-generated tracks uploaded every day, around 2 million per month, representing 44% of all new uploads. That’s a 650% jump in about 16 months compared with early 2025, when only around 10,000 daily tracks used AI. Yet listening tells a different story: these AI tracks account for only 1–3% of total streams on the platform. Deezer has also found that a large share of those AI streams are fraudulent plays, which it demonetises. For Malaysian listeners, this gap between supply and demand is key. The catalogue is being flooded with AI audio, but the typical user is still choosing familiar artists and songs, not randomly generated tracks sitting at the bottom of search results.
How recommendation algorithms keep AI in the background
Deezer doesn’t just track AI generated music; it also actively manages how you encounter it. The company labels AI tracks and filters them out of its music recommendation algorithms, meaning they won’t appear in its editorial playlists or in most personalised suggestions. That design choice helps explain why AI uploads form such a large share of the catalogue but only a sliver of listening time. For Malaysians using Spotify, Apple Music or YouTube Music, the situation is less transparent. Deezer’s CEO has urged the wider industry to adopt clearer labelling and anti-fraud measures, but other platforms have not disclosed comparable figures or detailed policies. Still, their music recommendation algorithms are built to maximise engagement and retention, which usually means pushing proven human artists, local favourites and major-label releases ahead of untested AI songs that risk listener fatigue or distrust.
Curiosity vs fatigue: why human music still wins
Listener behaviour suggests that AI songs on Spotify or any service face a trust and connection gap. Some people click out of curiosity when they see an AI tag or viral track, but repeat listening tends to favour human artists with stories, personalities and communities. Deezer’s tiny 1–3% streaming share for AI uploads, despite huge volume, reflects this imbalance. Many listeners treat AI tracks like novelty content rather than core listening. There’s also a growing sense of fatigue: with so much low-effort AI audio being uploaded, it becomes harder to find genuinely creative uses of the technology. For Malaysian users, daily listening is still dominated by artist‑driven songs in Bahasa Malaysia, English, Mandarin and Tamil, whether that’s local rock, K-pop, EDM or nasyid. AI tools may help make or remix these tracks, but audiences primarily connect with the artists, not the algorithms.
What Malaysian listeners can do now — and what comes next
If you want less AI in your playlists, start by checking track credits and descriptions where platforms label AI generated music, and lean on trusted editorial or human‑curated playlists rather than generic algorithmic mixes. Following favourite Malaysian artists and labels also nudges recommendation systems toward similar human-made tracks. If you’re curious about AI music, look for clearly tagged releases or AI-artist ‘profiles’ on services that support them, and experiment with radio or discovery playlists that highlight new creators. Looking ahead, Deezer-style labelling and stricter fraud controls are likely to spread, making it easier to filter or embrace AI content. Better rules could benefit indie musicians in Malaysia by reducing payment dilution from spammy uploads, while giving casual listeners more transparent control over how much algorithm-made music appears in their daily soundtrack.