How an innovative tech startup is helping music transcend language barriers

Though English was once the lingua franca of pop music, the genre’s evolution in the past decade has challenged its monopoly. Appetites have never been so borderless. The cultural and commercial value of Spanish, Korean and Chinese-speaking artists have continually broken glass ceilings.

Historically, international artists would have to ask or potentially pay large sums for an English or American star to record a verse in order to realise their global ambitions – but now, artists are reaching record-breaking levels of success without compromising on their cultural identity or relying on anyone but themselves. Puerto Rican star Bad Bunny has been crowned Spotify’s most-streamed artist globally for an unprecedented fourth time with 19.8 billion streams this year, and has been announced as the headliner for the Super Bowl halftime show in 2026. Global pop stardom has become a revolutionary act; there is no longer a hegemonic, cookie-cutter outline for success.

The tide is turning, but language continues to impose a limitation on an artist’s global reach. There’s an irony in the break-out phenomenon of KATSEYE – a “global girl group” – who still perform entirely in English. Nothing is sweeter to the ear than our mother tongue, after all – 80% of songs are released in a single language – and despite the fact that rhythm can be felt universally there is a depth of understanding in the lyrics which can unlock a whole new level of connection in a listener.

Though artists have found unprecedented levels of success without bending to the whims of the US and UK markets, language is still the iron maiden which continues to confine artists to their home audience. Imagine if a song’s story was never lost; imagine if an artist’s voice could not only be felt but understood without losing anything – tone, character, timing or emotion – in its translation.

It doesn’t have to be a pipe dream. Tech startup Moozz has found a way to close the gap between the demand and our reality. The small but devoted start-up of four founders recognised how uniting their decades of experience in music and tech could make an artist’s reach limitless. They have developed an AI model which translates songs into new languages while preserving the original features of their voice we had fallen in love with. The goal, they say, is to make music border-defying and empower an artist to reach a global audience without compromising on their artistic integrity.

Jérôme Keff is the former Managing Director of BMG Production Music France with over 20 years in publishing. He notes that the vast majority of independent artists are making very little money, many of whom are unable to pursue music full-time due to the low payouts from streaming and the losses made from touring, but the financial rewards of sync can be game-changing for their career. It can create passive, long-term income from a single track for years. If their music can adapt to the requirements of film and television regardless of the territory, chosen on merit rather than language, it empowers the artist to deservedly monetise their output.

The current state of affairs demands that if you wish to adapt songs into multiple languages, there is no other way around it other than to pay for them to be entirely re-recorded. Eduardo Garcia, who has previously scaled educational tech companies on an international level, illustrates the point with the example of Disney’s Frozen franchise and its inescapable soundtrack. They translated it into 41 languages and spent over $1.5 million on 1300 recording sessions which required hundreds of new, international vocalists. “With our technology, you could solve this problem in minutes at a fraction of the cost,” he says.

As soon as the subject of AI even brushes shoulders with music, there’s an understandable and instinctive resistance. It has become synonymous with artistic theft and being the enemy of the authenticity for which music has always been a refuge. But Moozz assure that theirs is not a large, general-purpose model – it’s small and artist-specific, trained case by case and only by the material provided or approved by the rightsholder. Each artist’s voice is processed independently to preserve that authenticity that is so highly prized.

To train the model, the artist must deliver approximately five songs. From this data set, the model has enough material from the voice to adapt it for another language. “Only with the agreement of the artist can all this work be done,” notes Keff. “We’re not just taking stuff from the internet. The ethics of this are really important to us, and we’re aware of how AI has been abused. It’s not the way we want to work. We want to work very closely with the artist and make sure all the rightsholders are respected.”

The rights and metadata for every translation always stay with the original creators. Moozz doesn’t claim ownership over the music it processes; it functions only as a tool which helps artists and labels localise their music to reach specific audiences. It’s still in its infancy, though they are in conversation with a major streaming platform and have had early pilots of the technology with labels, publishers and distributors. Now, their mission is to connect with the artists who can make this their responsible, human-focused vision for the future of music with AI a reality. Moozz wants to extend an invitation to artists and their teams to collaborate with them in shaping this emerging frontier of music reaching its universal potential.

“We want to work closely with the artist so that we can do something great with their voice without destroying the spirit, style or emotion of their original voice,” Keff says. “The technology is quite challenging – and I wouldn’t have said the same thing 20 years ago – but now we’re in a place where we can do something great and support artists in their ambition to reach the whole world.”