Deezer Finds Most Listeners Can’t Distinguish AI‑Generated Music
Key Points
- Deezer and Ipsos surveyed thousands of listeners on AI‑generated vs. human music.
- Only one participant correctly identified all three AI tracks; overall AI detection was about 43 percent.
- 71 percent of respondents were surprised by their inability to tell the difference.
- 80 percent want clear labeling of AI‑generated songs for transparency.
- Deezer automatically labels AI tracks from models like Suno and Udio and excludes them from recommendations.
- Spotify plans a credits system for AI disclosure but does not require mandatory labeling.
- Over 50,000 AI‑generated tracks are uploaded to Deezer daily, making up more than 34 percent of new additions.
- AI tracks account for roughly 0.5 percent of total streams, most of which are considered fraudulent.
- Artists and industry leaders view AI as a tool to augment, not replace, human creativity.
Deezer’s recent experiment, conducted with research firm Ipsos, revealed that the vast majority of listeners struggle to tell AI‑generated tracks from human‑made songs. In a survey of thousands of participants, only a tiny fraction correctly identified all AI songs, while many expressed surprise and discomfort at the results. The findings have sparked debate over transparency, labeling, and the potential impact of AI on the music industry, prompting Deezer and other platforms to consider new policies for AI‑generated content.
Deezer’s AI Music Experiment
Deezer partnered with research firm Ipsos to test whether listeners could differentiate between fully AI‑generated music and human‑made tracks. The study asked participants to listen to three songs and guess which, if any, were produced entirely by artificial intelligence. Participants who failed to correctly identify all three were counted as unable to tell the difference.
Survey Results Highlight Low Identification Rates
The survey involved thousands of participants. Only one person correctly identified all three AI tracks, while the overall success rate for recognizing AI content was roughly 43 percent. A large majority of respondents—71 percent—were surprised by their performance, and 51 percent reported feeling uncomfortable about not being able to distinguish AI‑generated music.
Public Sentiment on AI‑Generated Music
Opinions among listeners were mixed. About half of those surveyed (51 percent) believed AI would lead to “more low‑quality, generic‑sounding” music, yet only 40 percent said they would skip AI music if they knew it was AI‑generated. Concerns about the livelihood of musicians were prevalent, with 70 percent viewing AI songs as a potential threat, and 64 percent fearing a reduction in creativity.
Calls for Transparency and Labeling
Transparency emerged as a key demand, with 80 percent of participants wanting clear labeling of AI‑generated tracks. Deezer has responded by creating a system that automatically detects and labels AI content from popular models such as Suno and Udio, and it excludes labeled AI tracks from its algorithmic recommendations. Spotify, meanwhile, announced policies to combat AI‑related spam and impersonation, but stopped short of mandatory labeling, instead proposing a standardized credits system that relies on artists’ disclosure.
Industry Perspectives on Hybrid Content
Manuel Moussallam, director of research at Deezer, described hybrid content—music that incorporates AI elements—as a transparency and ethical issue rather than a technical one. He emphasized that responsible action from creators, distribution services, and streaming platforms is essential.
Volume of AI‑Generated Music on Platforms
Deezer reports receiving over 50,000 AI‑generated tracks per day, accounting for more than 34 percent of new music added to the service. Despite this volume, AI tracks represent only about 0.5 percent of total streams, and most of that fraction is considered fraudulent. Moussallam noted that while the sheer number of AI uploads poses operational challenges, it has not dramatically altered the listening experience for most users.
Artist Views on AI’s Role in Music
Artist Holly Herndon, who frequently uses custom AI models, echoed the sentiment that the mere ability to produce polished, AI‑generated songs does not guarantee audience interest. She suggested that artistic practice remains more sophisticated than simple AI output.
Future Outlook
While concerns about AI’s impact persist, Moussallam expressed optimism, stating that the industry is not moving toward a future where humans are removed from the creative process. Instead, AI is expected to become integrated into existing workflows, enhancing rather than replacing human creativity.