SZA Drags AI Manipulators Who Trained on 238 of Her Songs Without Permission

SZA is speaking out after discovering that artificial intelligence systems had allegedly trained on 238 of her songs, including tracks she believes may never have been released to the public. Posts shared to Instagram showed the Grammy-winning singer condemning the practice and the musicians who support it.

Newly launched AI-dataset detection technology now allows performers to search databases used to train AI music generators and determine whether their work appears in those collections. As reported by The Atlantic, findings from the project showed music from major stars and independent performers alike had been included in datasets containing more than 21 million songs.

Instagram Stories posted by SZA revealed anger directed not only at AI companies but also at musicians who support the practice. She described those backing unauthorized AI training as "disgusting" and suggested no justification could excuse the use of copyrighted material without permission.

Subsequent messages expanded the dispute beyond copyright concerns. SZA argued that Black musicians are disproportionately affected by AI scraping and urged creators not to allow their work to become fuel for systems built without permission or compensation.

Producer Diplo was also mentioned in posts that circulated online after SZA reacted to the findings. Her comments added fresh attention to an increasingly heated clash between creators and technology companies developing AI music tools.

Unauthorized use of copyrighted recordings remains a major fault line. Labels, technology firms and performers continue to disagree over whether copyrighted recordings can be used to build AI systems without direct approval from the people who created them.

Suno and Udio, two prominent companies in the AI music space, have already faced lawsuits tied to the use of copyrighted music. SZA's criticism adds another high-profile voice to a battle that continues to reshape conversations about ownership, compensation and creative control in the AI age.