SZA Urges Black Musicians to Reject AI After Learning Model Used 238 of Her Songs
SZA is once again making it clear that she wants no part in artificial intelligence taking from artists without permission.
After The Atlantic released a new AI detection tool allowing musicians to search whether their songs appear in training datasets, the “Kill Bill” singer said she found 238 of her songs listed, including what she believes may be unreleased music.
“Jus checked and music AI has trained off 238 of my songs,” SZA wrote on Instagram Stories. “I’m certain some unreleased.”
Her frustration did not stop there. The 36-year-old artist also called out musicians who support the practice, writing that there is “NOTHING” anyone could say to make it acceptable.
The database, created by The Atlantic’s Alex Reisner, pulls from several large music datasets that include millions of songs linked through platforms such as Spotify and YouTube. Artists including Beyoncé, Taylor Swift, SZA and many independent musicians reportedly appear in the data, raising new questions about consent, ownership and whether creators have any real control once their work is scraped into an AI system.
Why Black Artists Say This Feels Familiar
For Black artists, the concern is deeper than new technology.
SZA also warned Black musicians and producers not to give away what she called their “vibranium,” pointing to the global influence of Black creativity despite Black Americans making up a smaller share of the U.S. population.
“We make up 13% of the American population yet influence the world w our sound and perspective. I AINT HEARD A WHITE AI SONG YET. We have no protection in legislature medical or creative. The easiest to steal from. DO NOT GIVE AWAY YOUR VIBRANIUM !!! DO NOT TRAIN AI W YOUR GENIUS. F— these weird a— vultures. I want smoke all summer.” SZA wrote.
Her message speaks to a long history in the music industry, Black artists shaping the sound of popular culture while others profit from their work, style and innovation. From blues and jazz to rock, hip-hop, R&B and pop, Black creativity has repeatedly been borrowed, repackaged and monetized.
Artificial intelligence now threatens to speed up that same pattern.
Instead of one artist being copied by another, AI systems can absorb thousands or millions of songs, voices, lyrics, cadences and production styles at once. That makes the issue less about one stolen hook and more about an entire creative identity being turned into raw material.
What Can SZA Actually Do?
The law around AI training and copyright remains unsettled. A song appearing in a dataset does not always prove which company used it, whether it was copied into a specific AI model or whether it directly generated an infringing output. That makes enforcement difficult, especially for artists without major label backing.
Still, artists do have options.
They can document where their work appears, save screenshots and metadata, contact their label, publisher or distributor, and ask whether legal notices or takedown demands have been sent. They can also push their teams to audit licensing agreements to make sure AI training rights have not been quietly included in broad contract language.
Major labels have already taken legal action. Sony Music, Universal Music Group and Warner Records filed copyright lawsuits against AI music companies Suno and Udio, accusing them of using copyrighted recordings without permission to train their platforms.
Because SZA’s catalog is tied to a major label system, she may have more legal leverage than independent artists whose songs appear in the same datasets.
How Creators Can Protect Their Work
For Black artists, writers, producers, influencers and content creators, protection starts with paperwork and proof.
Creators should register copyrights for music, lyrics, videos, photography, written work and other original content. They should keep dated files, project stems, drafts, contracts and uploads showing when work was created and published. When possible, creators should use watermarks, licensing terms and platform settings that make unauthorized scraping harder to ignore.
They should also read contracts carefully. Any agreement involving labels, management, distribution, brand partnerships or content platforms should be reviewed for language giving companies permission to use work for “machine learning,” “data mining,” “training,” “synthetic media” or “future technologies.”
Creators can also send takedown notices when AI-generated content copies their work too closely, join collective legal efforts, support artist unions and advocacy groups, and demand legislation requiring consent, credit and compensation before creative work is used to train AI.
Some artists and producers have embraced AI, arguing that the technology is inevitable. Others see it as another industry shift where creators are told to adapt after their work has already been taken.
That divide is growing.
For SZA, the issue is not simply that AI exists. It is that Black artists, who have already shaped so much of global culture, may again be asked to watch others profit from their genius without permission, payment or protection.
