Every language, not just a few
Most voice systems are built for a handful of widely spoken languages. We research how voice AI can serve the languages people across Africa actually speak, even where recorded data is scarce.
For many people, speaking is the most natural way to use technology. We research voice AI that understands and speaks the languages people use every day across Africa, and that works in the real places they use it.
One voice. Many languages.
Voice can make AI available to people who would rather speak than type, or who find typing hard. Yet most voice systems struggle with tonal languages, local accents and the way people mix languages in a single conversation.
What would voice AI look like if it were built for the languages and conditions of Africa from the start?
That is the question our Voice AI research exists to answer. What we learn shapes Sixteen, and the wider research we share.
Most voice systems are built for a handful of widely spoken languages. We research how voice AI can serve the languages people across Africa actually speak, even where recorded data is scarce.
Tones change meaning, accents vary from city to city, and many people move between languages mid-sentence. We study how voice AI can handle speech as it really sounds, in noisy, everyday places.
A voice can be cloned and recordings can be misused. We build consent, privacy and safeguards against misuse into voice AI from the start, guided by our safety and evaluation principles.
Coming soon
Sixteen is the next product from NuvoAI, and the first to come out of our Voice AI research. We will share more about it soon.