Suno CEO Mikey Shulman told The Guardian earlier this year that his own tool had been described to him as the “Ozempic” of the music industry. Everybody uses it, but nobody talks about it.

That strict “don’t ask, don’t tell” mentality is now widespread among top producers and songwriters, according to the Rolling Stone report. They are using the technology heavily but staying silent publicly out of fear of backlash. One warning example is singer Teddy Swims, who faced strong criticism after openly admitting to using AI.

Producer David Baron said in the report that he is convinced AI-generated music has already made its way onto the Billboard charts. Lauren Christy, a songwriter for Avril Lavigne and Britney Spears, summed it up bluntly: “That train has left the station.” The reason for the lack of transparency is that major labels do not have reliable software capable of detecting AI-generated music. Instead, the industry is relying on what is essentially an honor system.

More than half of new hip-hop samples may be AI-generated

The shift is particularly visible in hip-hop. Instead of licensing real soul records from the 1960s or 1970s or hiring studio musicians, producers are now using AI to generate fictional retro samples. Producer Young Guru, Jay-Z’s longtime engineer, estimates that “more than half” of sample-based hip-hop is now created this way.

The quality of AI-generated vocals has reached a level that even professionals find unsettling. Christy described a singer who, after hearing an AI demo, reacted with frustration: “I hate this robot. She sings it better than I do.”

The growing use of AI can also be measured. According to a Sonarworks survey of more than 1,100 producers, sound engineers, and songwriters, seven out of ten respondents experiment with AI tools at least occasionally, while one in five uses them regularly. Most apply the technology to narrowly defined, time-saving tasks such as audio restoration, stem separation, and mastering. Sonic matching — transferring the sound character of a reference track onto one’s own mix — now takes minutes instead of hours or even days thanks to AI.

Smaller players are hit the hardest

The speed at which AI can deliver finished demos is fundamentally changing workflows. According to the report, Christy received a text message from a “major star” looking for new songs. She fed her lyrics and chords into an AI system and quickly sent back a finished demo. The artist wanted to record the song.

While such scenarios create new opportunities for established songwriters, they are eroding the market for second-tier professionals. Session musicians who once sang demo vocals and studio assistants are losing work, according to Rolling Stone. The stock and production music market, for example for smaller TV productions, is also said to be effectively “finished.”

Songwriter Michelle Lewis, who has written for artists including Cher and Hilary Duff and co-founded the advocacy group Songwriters of North America, said that writers in Nashville and Los Angeles are using tools like Suno to generate fully arranged demos from lyrics and chords.

“In private, songwriters say, ‘This is pretty great,’” Lewis said. “You don’t have to split your copyright, you can write alone, and you don’t need to pay a producer.” Lewis, who also works in children’s animation, described that market as “easy prey” for AI replacement. Overall, she said, “nobody is working anymore.”

AI copyright issues remain a major fear

Uncertainty remains high around copyright. On one hand, it is still unclear whether the music generators themselves are lawful. Market leader Suno is currently involved in copyright disputes, though it has already announced plans to work more closely with the music industry later this year using even stronger models.

The recently released Suno 5.5 offers impressive quality and can now even incorporate a user’s own voice into songs. One Suno investor recently acknowledged that the music generator is in direct competition with human-made music. That point is particularly sensitive because it could be used in court as an argument against fair use.

Google already offers its own music generator and says that Lyria 3 was trained exclusively on content for which Google had permission. OpenAI is also reportedly expected to follow soon.

On the other hand, artists remain uncertain about whether AI-generated content — whether text, images, or audio — can be protected by copyright at all. An AI-generated song without copyright protection would ultimately have little value. With AI-generated samples, there is also the question of whether they meet the legal threshold of originality. So far, these cases have been handled individually, and most rulings have tended to go against AI.