This week, three songs generated by artificial intelligence have reached the top of music charts, securing spots at the pinnacle of both Spotify and Billboard rankings.

Walk My Walk made waves with Breaking Rust’s “Livin’ on Borrowed Time” leading the U.S. “Viral 50” chart, which highlights the “most viral songs” daily on Spotify. The anti-immigrant anthem “We Say No, No, No to an Asylum Center” by JW “Broken Veteran” also soared to the top of Spotify’s global viral chart during this timeframe. Additionally, “Breaking Rust” landed in the top five globally.

The lyrics of “Walk My Walk” include the line, “If you don’t like the way I talk, you can kick a rock,” directly addressing critics of AI-generated music.

Shortly after climbing the charts, the Dutch song vanished from Spotify and YouTube, along with all other tracks by Broken Veteran. Spotify told Dutch outlet NU.nl that it had not removed the music and that the rights holder was responsible. Broken Veteran expressed confusion over the disappearance, stating he is investigating and hopes for a resolution soon.

Opting to remain unnamed, Broken Veteran shared with the Guardian via email that he views AI as “just another tool for expression,” especially for individuals like himself who have important messages but lack formal musical training. He emphasized that the technology has “democratized music production” and clarified that his song critiques government policies, not immigrants.

For three weeks, “Walk My Walk” has led Billboard’s Country Digital Song Sales chart, which tracks downloads and digital purchases. This chart is considered minor compared to Billboard’s broader metrics like “Hot Country Songs” and “Top Country Albums.” Breaking Rust has yet to respond to media inquiries.

AI-Generated Music’s Growing Presence

These three tracks exemplify the surge of AI-generated music flooding streaming services. A recent study by Deezer reveals that roughly 50,000 AI-generated songs are uploaded daily, representing 34% of all music submissions.

“Walk My Walk” and “We Say No, No, No to an Asylum Center” are not the first AI tracks to gain popularity. Earlier in the summer, an AI-produced song by a group named Velvet Sundown achieved over 1 million streams on Spotify, which one of its members later referred to as “art quackery.”

Ed Newton Rex, a musician and founder of a nonprofit aiming to ensure fair data training for generative AI companies, notes that the high volume of AI-generated songs available online is a significant factor behind the emergence of AI hits.

“This reflects a trend of rapidly growing interest in AI music, driven primarily by the volume of content,” he explained. “Daily, we see 50,000 new songs competing with human artists, marking the rise of a new, highly scalable competitor built through exploitation.”

AI music quality has noticeably improved since its early days. In a survey conducted as part of the study, Deezer found that 97% of the 9,000 participants from eight countries could not differentiate between AI-generated music and human-created compositions.

“This is undeniable. It’s now fairly safe to say that the top-tier AI music is indistinguishable from human-composed tracks,” Newton-Rex stated.

Emerging Distribution Models for AI-Generated Music

The success of AI-generated tracks on Spotify transcends mere quality. Much like various domains within the AI economy, numerous tools and platforms facilitate the distribution of AI music, along with user sub-communities eager to share strategies for navigating the system.

Jack Righteous, a blogger focused on AI content creation, has highlighted how his followers can generate a “passive income stream” through a music distribution service called DistroKid, which allocates royalty fees to creators whenever an AI track is streamed on platforms like YouTube, Spotify, or TikTok.

DistroKid is part of a wider ecosystem of online music distribution services, including Amuse, Landr, and CD Baby, all helping creators publish their music on major platforms like YouTube and Spotify. These services have varying policies regarding AI-generated content, with blogs like noting that DistroKid is “more forgiving.” This includes some hits from Breaking Rust such as “Livin’ on Borrowed Time” and “Resilient,” which appear to be distributed by DistroKid.

“In essence, most AI music you encounter isn’t being handled by a legitimate label. It’s crafted by individuals in their personal spaces and uploaded to distribution platforms,” said Chris Dalla Riva, author of “Unknown Territory,” which delves into the data behind music virality.

When approached for comments, Spotify cited their policy regarding AI-generated tracks.