Most listeners can’t distinguish between music created by humans and music generated entirely by artificial intelligence (AI), according to recent research. A study by streaming service Deezer and Ipsos found that 97% of people struggle to identify AI-generated tracks when compared to human-made music. This raises critical questions about the future of music creation, consumption, and the value of artistic authenticity.
The Experiment and Its Nuances
Deezer’s initial experiment involved 9,000 participants who listened to three tracks—some made by AI, some by humans—and were asked to identify the AI-generated ones. The threshold for failure was high: missing even one correct identification meant the participant was deemed unable to distinguish between the two.
However, when analyzed differently, the results were less dramatic. Participants correctly identified AI-generated tracks 43% of the time if their responses weren’t grouped as a collective failure. Some even intentionally guessed incorrectly, believing a particularly poor-sounding track must be human-made as part of a trick.
Transparency Concerns and Industry Response
The inability to differentiate between AI and human music left 71% of participants surprised and 51% uncomfortable. Over half (51%) fear that AI will flood the market with low-quality, generic content. Yet, surprisingly, only 40% said they would automatically skip AI music if they knew its origin.
This raises a major issue: transparency. A resounding 80% of people want AI-generated music to be clearly labeled. Deezer is taking steps to address this, developing a system to automatically detect and label content created by models like Suno and Udio, excluding it from algorithmic recommendations.
Spotify is also responding but with a more cautious approach. It’s implementing policies to fight AI-driven spam and impersonation, but refuses to commit to blanket labeling of AI music. Instead, it proposes a “nuanced” credit system relying on artists and labels to self-disclose AI usage.
The Scale of the Problem and Potential Solutions
The sheer volume of AI-generated music being uploaded is staggering. Deezer receives over 50,000 AI tracks daily, accounting for 34% of all music added. Despite this influx, AI music currently represents only 0.5% of total streams—mostly fraudulent.
Experts emphasize that the core problem isn’t technical, but ethical. Manuel Moussallam, Deezer’s research director, argues that transparency and responsible action from creators, distributors, and platforms are crucial.
The Future of Music Creation
While concerns about AI replacing human artists are valid, many remain optimistic. Holly Herndon, an artist who uses AI in her work, believes that “just because anyone can create polished kitsch doesn’t mean anyone will care about those songs. ”
The survey showed that 70% of respondents fear AI will harm musicians’ livelihoods, and 64% believe it could stifle creativity. However, Moussallam suggests AI will likely integrate into creative processes rather than eliminate human involvement.
Ultimately, the rise of AI-generated music forces a reckoning with the value of authenticity, the need for clear labeling, and the evolving relationship between humans and machines in the world of art. The question is no longer whether AI can create music, but how we ensure it doesn’t undermine the integrity of the industry.
