SZA Discovers 238 of Her Songs Used to Train AI Music Model

Grammy-winning artist SZA has intensified her opposition to artificial intelligence in music after discovering that 238 of her songs—including unreleased tracks—were used to train AI models without her consent.

The singer-songwriter shared her discovery on Instagram Story after searching an AI music database for her name. The revelation prompted her to issue a direct appeal to fellow musicians, particularly Black artists, to refuse participation in AI training systems.

Discovery reveals scope of unauthorized use

SZA's search revealed that hundreds of her compositions, both released and unreleased, had been incorporated into AI training datasets. The finding adds concrete evidence to ongoing debates about how generative AI companies source their training material.

In her public statement, she condemned musicians who support AI training practices. In a separate post to her private Instagram account, she specifically criticized producer Diplo, claiming he holds equity in AI music startup Suno and is "actively attempting to train it on the best and brightest Black minds of writers and producers," according to Rolling Stone, which first reported the details.

While Diplo has no confirmed public connection to Suno, the Wall Street Journal identified him as an investor in Aaru, a billion-dollar AI startup, in March 2026.

Legal battles over AI training continue

Suno ranks among the most widely used generative AI music services globally. In 2024, three major record labels filed suit against both Suno and competitor Udio, alleging copyright infringement. Suno has defended its practices, arguing that using copyrighted music for model training falls under fair-use doctrine.

The platform initially allowed users to generate songs from text prompts. Subsequent updates introduced capabilities to create music from uploaded videos, photos, partial compositions, and vocal recordings.

Why it matters

SZA's call to action highlights a critical tension in AI development: the disproportionate use of Black artists' creative output to train systems that may ultimately compete with or devalue their work. Her argument that Black musicians "make up 13% of the American population yet influence the world" with their sound underscores concerns about who benefits from AI systems built on minority artists' contributions. As generative AI becomes more sophisticated, questions about consent, compensation, and cultural exploitation in training datasets will only intensify.

Artist frames issue as cultural exploitation

SZA framed the issue in terms of systemic vulnerability, noting that Black artists lack adequate legal protections for their creative work. "We have no protection in legislature medical or creative. The easiest to steal from," she wrote, urging artists not to "give away your vibranium"—a reference to the fictional resource from Black Panther.

This isn't SZA's first critique of AI technology. She has previously connected AI development to environmental racism, pointing to pollution and water consumption in underserved communities caused by data centers powering AI systems.

Diplo responded to AI criticism in April 2026, stating on X that creatives must "adapt or just like give up and become an uber driver." He has faced recurring accusations of cultural appropriation throughout his career. In a 2018 Guardian interview, he acknowledged his privilege as "a middle-class white American" and called criticism "a small price to pay."

Rolling Stone first reported the details of SZA's statements and the broader context of AI music litigation.