Roughly two‑thirds of U.S. consumers say they are not particularly interested in listening to music that is partly or entirely created using artificial intelligence, according to a new survey by Bain & Company.
Bain’s Media Consumption Survey 2025 asked several thousand people about their interest in engaging with a variety of forms of entertainment and media, from video games to social media to books. Researchers found that U.S. consumers have the lowest tolerance for AI when it is used to make music, news and magazines and books. Roughly 62% of respondents said they would not or were less likely to engage with AI-generated music, with another 30 percent of respondents saying they were equally likely to engage with AI-generated music.
Related
“That was one of the strongest endorsements of human-created content across media types,” Bain researchers Matt Keith and Nicole Magoon wrote.
Consumers were even less warm to AI-generated news, magazines and books, with 65% and 71% saying they would not engage or were less likely to engage with AI-generated content in those mediums, respectively.
Nonetheless, most respondents were open to AI being used in limited ways, including to enhance vocals or instruments on a track or improve lyrics.
More than half of the study’s some 3,500 respondents said they were equally interested or more interested in hearing songs that used AI to write lyrics, and 56% of respondents expressed equal or greater interest in songs where AI was used to translate lyrics into another language and replicate the singer’s voice singing the lyrics in that other language.
Around 60% of respondents expressed equal or greater interest in songs that used AI to create instrument sounds or to enhance the sounds of instruments or vocals, and 67% — two-thirds — said they were equally or more interested in songs that used AI to help with the initial idea, development or to improve lyrics.
Researchers Keith and Magoon say these findings reflect an audience that is mostly ready to accept AI as a new additive tool in music so long as it doesn’t replace artists’ authenticity.
They add that AI presents needed opportunities for the major music companies, creators and companies in the live entertainment space.
Existing creators can use it to “work faster [and] experiment more” to meet the need to release “more music more frequently,” angling for better exposure across streaming platforms’ algorithms and playlists and stand out from the new class of hobbyist creators enabled by AI.
The researchers suggested that labels and live music companies could use AI to improve and extend older artists’ ability to perform live by using AI tools for “instant voice de-aging” and improving sound mixing to match a variety of venues and accoustics.
“The creators and companies that accept [generative AI] and move swiftly to capitalize will have a shot at enduring success,” the authors wrote, adding they suggest the industry soundtrack these next steps to R.E.M.’s “It’s the End of the World as We Know It (And I Feel Fine).”
