A new global survey from Deezer has confirmed a growing suspicion: most listeners can’t tell the difference between an AI-made track and the real deal.
According to the study, 97% of listeners cannot reliably tell the difference between fully AI-generated music and tracks made by human artists. The results paint a picture of a public that’s curious about AI but increasingly uneasy about its negative impact on music culture and the people who make it.
The streaming platform surveyed 9,000 participants across eight countries to listen to three tracks and identify which ones were AI-generated. More than half of respondents said the results made them “uncomfortable” and only 19% listed trust among their feelings toward artificial intelligence.
The sheer volume of AI on streaming platforms keeps rising as well. Deezer reports that roughly 50,000 fully synthetic tracks are now delivered to its platform every day, representing more than a third of daily uploads. This follows warning signs the company shared earlier this year, when it described a surge of AI uploads and the rollout of its early detection tools.
Listeners are also split on whether AI will help or harm the music they consume. Nearly half believe AI could improve music discovery, but most worry it will flood streaming platforms with low quality and generic releases. Almost two thirds believe AI could lead to a decline in creativity overall. There’s also a shift in listening behavior: 40% of streaming users say they would skip a fully AI-generated track the moment they realize it is not made by an actual artist.
The broader industry concerns highlighted in the survey are serious. A study from CISAC and PMP Strategy estimates that nearly a quarter of creators’ revenue could disappear by 2028 as AI-generated output increases.
Deezer is currently the only major streaming service that tags fully AI generated tracks for users. It also removes these tracks from all algorithmic recommendations and editorial playlists, a move designed to keep AI content from diluting the royalty pool. The company claims that a large share of streams on fully AI generated tracks appear to be fraudulent, and any suspicious activity is excluded from payouts.
You can read the full study here.
The post 97% of Listeners Can’t Tell the Difference Between AI Music and Human-Made Tracks appeared first on EDM.
Recommended Comments