AI music floods streaming services, but listeners aren’t buying it
▼ Summary
– Generative AI music began as an experimental gimmick, but platforms like Suno and Udio made it accessible to anyone, leading to a flood of machine-made tracks on streaming services.
– By 2025, Deezer reported that AI-generated content made up 34% of uploads, growing to 75,000 daily tracks, while Spotify removed over 75 million spam tracks in a year.
– Streaming platforms like Deezer, Qobuz, Apple Music, and Spotify have implemented detection or voluntary labeling systems, but enforcement varies, with Apple relying on self-reporting and Spotify using AI credits.
– Public opinion is largely negative, with 66% of people never knowingly listening to AI music and 52% unwilling to listen to favorite artists if AI was involved, though only Bandcamp has outright banned it.
– Despite the surge in AI uploads, consumption remains low, accounting for only 1% of streams on Deezer, with high fraud rates suggesting waning interest, yet companies expect AI to become a standard tool in the industry.
The explosion of AI-generated music on streaming platforms has turned into a double-edged sword for the industry. While the technology has democratized music creation, allowing anyone with a prompt to produce a track, the sheer volume of machine-made content is overwhelming services and frustrating listeners. A growing number of users and artists are demanding action, but the streaming giants are struggling to find a consistent, effective response.
What began as a novelty in 2018 and 2019, with experimental albums like I AM AI and Proto, has become a tidal wave. The launch of user-friendly tools like Suno and Udio in late 2023 and early 2024 opened the floodgates. By September 2025, Deezer reported that 28 percent of its uploads were fully AI-generated. That figure jumped to 34 percent by year’s end, with over 50,000 synthetic tracks hitting the platform daily. The problem is only accelerating, with Deezer now seeing 75,000 AI uploads per day, threatening to eclipse human-made music entirely. Spotify, meanwhile, removed over 75 million spam tracks in a single year.
Deezer has been the most proactive, becoming the first major platform to deploy a detection and labeling system. It now demonetizes 85 percent of AI streams and prevents its algorithm from recommending them. CEO Alexis Lanternier stated that “AI-generated music is now far from a marginal phenomenon,” urging the entire ecosystem to act to protect artists’ rights. Qobuz followed with its own detection system and an AI charter, vowing to keep its editorial and curation work human-centric. Apple Music also introduced labeling, but its system relies on voluntary self-reporting through “Transparency Tags,” with no clear enforcement mechanism. Spotify has taken a similar voluntary approach, recently launching “AI credits” that let artists specify how AI was used, whether for lyrics, vocals, or instrumentation. It is working with the standards group DDEX to create an industry standard, but not all major players are on board yet.
Public sentiment is overwhelmingly negative. A Deezer and Ipsos study found that 51 percent of people believe AI will lead to more low-quality, generic music. A separate poll by The Hollywood Reporter and the Frost School of Music revealed that 66 percent of listeners never knowingly stream AI-generated music, and 52 percent would reject even their favorite artist’s work if it involved AI. Researchers from Singapore attribute this bias to AI’s perceived lack of expressive intent, arguing that it cannot forge the emotional connections central to human music.
Despite this backlash, only Bandcamp has outright banned AI-generated music. Its policy prohibits music “generated wholly or in substantial part by AI,” but enforcement relies entirely on user reports, not proactive scanning. Every other major platform remains hesitant to take such a hard line, largely because they expect AI to become a standard creative tool. As Spotify’s Sam Duboff noted, “the use of AI in music will increasingly be a spectrum, not a binary.”
The flood shows no signs of slowing. Deezer’s research director, Manuel Moussallam, confirms that AI uploads will likely keep increasing. However, there is a potential silver lining: consumption of AI music is not keeping pace. While uploads have grown by nearly 40 percent, AI tracks account for only about 1 percent of streams on Deezer, up from 0.5 percent in November. The fraudulent stream rate for AI music has actually jumped from 70 percent to 85 percent, suggesting listeners are actively avoiding it. The novelty appears to have worn off.
Creations like Velvet Sundown and Solomon Ray may generate buzz for being AI, but not for musical quality. The real threat is to working musicians, session artists, and library music composers. Yet AI is increasingly being adopted behind the scenes, from songwriting in Nashville to replacing sampling for hip-hop producers. Diplo has even warned creatives to adapt or risk being left behind.
The industry is caught in a bind. Companies are reluctant to penalize AI use because they foresee it becoming standard. Even Spotify’s “Verified by Spotify” badge, designed to guarantee a human behind a profile, leaves the door open for AI acts, calling “artist authenticity” a complex and evolving concept. But with Suno users generating an entire Spotify’s worth of AI slop every two weeks, pressure for dramatic action is mounting. The Deezer/Ipsos study found that 45 percent of people want the ability to filter out all AI-generated music from their libraries. No streaming service has committed to this, and implementing it would require a universally adopted labeling standard and reliable detection tools. If a simple switch existed to hide all generative AI music, a significant portion of listeners would likely flip it without hesitation.
(Source: The Verge)
