When the voice of a salsa legend is resurrected through silicon, the result is a delicate blend of homage and caution.

The estate of Cuban icon Celia Cruz has confirmed that her voice has been cloned by ElevenLabs, marking the first time a Latin artist’s vocal likeness has been recreated with artificial‑intelligence technology. The model will be applied only to educational content, book narration, and interactive projects that draw directly on the singer’s own words. According to the estate, the voice will be available exclusively in Spanish, and any English use will be limited to brief excerpts that preserve authenticity.

Executor Omer Pardillo told the Spanish news agency EFE that the voice has already been registered and that the estate will retain “absolute control” over its deployment. He emphasized that the project is intentionally narrow, stating, “politics is off limits.” The statement reflects Cruz’s deep ties to the Cuban exile experience and the political memories that shaped her music.

ElevenLabs, the company that built the model, has previously cloned the voice of poet Maya Angelou. In a statement, the firm said the collaboration with Cruz’s estate is meant to carry her voice into a new technological chapter “in an intentional and dignified way.” Bridget Ferris, ElevenLabs’ head of talent partnerships, added that Cruz’s energy, joy, and cultural impact are unique and deserve careful handling.

The decision follows a long history of preserving the legend’s legacy. The Smithsonian Institution holds her dress and shoes, the U.S. Postal Service issued a commemorative stamp, and the Treasury announced that she would be the first Afro‑Latina to appear on U.S. currency. In 2026, Cruz was posthumously inducted into the Rock & Roll Hall of Fame.

While the model can replicate accents, timbre, and phrasing, the estate acknowledges that it cannot convey the exile experience or the communal release that defined her performances. Pardillo said the voice is not the whole woman and that careless use could reduce her legacy to a puppet. He pledged that the estate would pursue legal action against any unauthorized use and stressed that the voice would never be paired with content Cruz would not have approved.

The project serves as a test case for Latin music’s future. It demonstrates how AI can preserve cultural memory while raising questions about control, authenticity, and potential misuse. By imposing strict guardrails—limiting usage to non‑political, educational, and narrative contexts—the estate offers a cautious model that other artists’ estates might follow.

As AI voice cloning spreads across entertainment, the Celia Cruz initiative shows that legacy holders can harness the technology while protecting the integrity of the artist’s image. The voice remains a guarded echo, available only for projects that honor the singer’s spirit and the history she embodied.

In the coming months, the estate will monitor the voice’s deployment and may broaden its use to new educational programs. The outcome of this experiment could influence how other estates approach AI‑based legacy preservation.

The voice stands as a controlled, dignified tribute to a Cuban exile icon—a reminder that technology can both preserve and commodify cultural heritage.