Real-time AI translation: erasing the language barrier

🇬🇧 English | 🇫🇷 Français

I came up with the idea for Caakus at 15, in Cameroon, the year I sat my baccalaureate. At the time it was only a hunch: what if you could talk to anyone, anywhere in the world, as easily as you talk to your neighbour?

At 17, I arrived in Germany on my own. I learned German up to C1 level, then began studying computer science. And I understood, from the inside, what it means to live in a language that is not your own: every conversation takes effort, and every misunderstanding can close a door.

That experience convinced me of one thing: language is still the biggest barrier between human beings today. Bigger than distance, bigger than the cost of travel. The internet has connected the world’s computers, but it has not yet truly connected people.

A connected world… that is still compartmentalised

We have never been so connected. Yet most of our online exchanges remain locked within our own language and our own country. A student in Dakar rarely talks live with a student in Seoul. A trader in Abidjan struggles to deal with a craftsman in São Paulo. Not for lack of interest, but because they do not share a common language.

Social networks show us the world, but we look at it from afar. Dating apps remain largely local. Random video chat sites, for their part, long stood for shocking content and abuse, to the point of driving away most users acting in good faith.

What changes everything: real-time translation

Artificial intelligence can now do something that was still science fiction five years ago: listen to a person speaking in their own language and render their words in another language, almost instantly, with live subtitles. Two people with no language in common can now understand each other during a video call.

This is a profound change. For centuries, communicating with the outside world required years of language learning, or an interpreter. Tomorrow it will be as natural as switching on your camera. For Africa, where hundreds of languages live side by side and where much of the youth is connected through the smartphone, this shift opens up huge prospects: cultural exchange, learning, professional opportunities, and a diaspora that stays connected to home.

What I learned building Caakus

I built Caakus alone, from A to Z: a platform where you pick a country on a 3D globe and, within seconds, find yourself on a video call with someone who lives there, straight from your browser, without downloading anything. AI translation lets people understand each other even without a common language.

Three lessons seem to me to reach well beyond my own project.

First, the desire to connect is universal. Without any marketing budget, Caakus has passed 48,000 users in more than 192 countries, who have already made more than 166,000 live video calls. Nine calls out of ten connect two different countries. People are not just looking to talk: they are looking to talk with somewhere else. The need was already there; the tool was missing.

Second, trust is the precondition for everything. A platform open to the world can only succeed if everyone feels safe on it. That is why we chose a protective AI that runs directly on the user’s device: it instantly blurs inappropriate content, without the video having to go through a server. Technology must protect before it entertains.

Third, a useful product spreads by itself. When an experience meets a real need, users become its best ambassadors. It is a valuable lesson for African entrepreneurs, who often face tight budgets: growth depends first not on the money spent on advertising, but on the value created for people.

Challenges not to be underestimated

It would be naive to present machine translation as a perfect solution. Cultural nuances, humour and local expressions remain hard to convey. African languages in particular are still poorly represented in the data used to train AI models. This is a major issue: if we do not make sure that Wolof, Swahili, Yoruba, Lingala or Ewondo are taken into account, Africa risks being a spectator of a revolution that should benefit it first.

There is also the question of privacy. A technology that listens to our conversations must be designed with strict respect for personal data. User trust cannot be decreed; it is built through transparent and demanding technical choices.

An opportunity for a generation

I believe African youth has a central role to play in this movement, not only as users but as creators. Artificial intelligence tools are now within reach of a motivated student with a computer and a connection. It has never been more possible to build, from Dakar, Douala, Kigali or Worms, a product used around the world.

The history of human communication has been shaped by a few great breakthroughs: writing, the printing press, the telephone, the internet. Real-time AI translation could be the next. For the first time, the language we speak will no longer determine who we can talk to.

My wish is simple: that one day a young person from any village can, within seconds, meet someone on the other side of the world and understand each other. Not from afar, not through a frozen screen, but in a real conversation. That, in the end, is what erasing borders means.

Bryan Lewis Dongue Ndiffo is the founder of Caakus, a live video calling platform for meeting people around the world, with real-time AI translation. A student of applied computer science at Hochschule Worms in Germany, he came up with the idea for Caakus at 15, in Cameroon.

À lire aussi

Traduction en temps réel par l’IA : effacer la frontière de la langueLFR 2026 rejetée par la majorité Pastef : analyse politique, budgétaire et institutionnelle des conséquences pour le SénégalSénégal : le rejet de la loi de finances rectificative, un rendez-vous manqué avec la responsabilité parlementaire