AI and digital: what was said at the 2nd Francophone Forum on Digital and AI Governance

At SAYNA, we keep a close eye on what the Organisation internationale de la Francophonie (OIF) is doing. It’s a partner we’ve already worked with in Madagascar through the D-CLIC program, which trains young people for digital jobs across the Francophone world. So when the OIF organizes the 2nd Francophone Forum on Digital and AI Governance, we listen — and we bring you the rundown.

The event took place on July 3, 2026 at the Graduate Institute in Geneva, with more than 900 registrants — three times more than the first edition in 2025. The timing was no accident: the forum ran three days before the UN’s very first Global Dialogue on AI Governance, the AI for Good summit, and the WSIS+20 forum, all in Geneva. The stated goal: to get a coordinated Francophone voice heard in international debates largely dominated by English.

On the mic: ITU Secretary-General Doreen Bogdan-Martin, UN Tech Envoy Amandeep Singh Gill, Smart Africa Director General Lacina Koné, Ambassador Henri Verdier, and representatives from ICANN, UNESCO, Afnic, Oracle, Québec, Senegal, Congo… Eight hours of debate, three big themes. Here’s the rundown.

1. A sobering fact: AI is highly concentrated

The number was dropped right at the opening: 75% of the world’s AI compute power is in the United States, 15% in China. The rest of the world shares what’s left. And on the data side, only 2% of African data is hosted on the African continent, compared to roughly 80% of European data hosted in Europe.

Another imbalance flagged by Henri Verdier: about 92% of the training data behind large AI models is in English. The consequence: these models carry only a fraction of human experience — along with their cultural and linguistic biases.

Amandeep Singh Gill summed up the stakes in one striking line: “The risk isn’t just technical, it’s civilizational: the risk of an intelligence that mistakes one humanity for all of humanity.” He then threw down a challenge to the Francophonie: become a “living lab” for multilingual AI — open corpora, shared data, and above all talent trained on all five continents.

2. Africa isn’t behind — it’s the new center

This was probably the talk that resonated with us the most: Lacina Koné, Director General of Smart Africa. A few of his lines deserve to be quoted as is:

“Africa isn’t behind, it isn’t racing anyone.”

“There’s no such thing as the ‘Global South’: Africa is the new center.”

“AI governance doesn’t start with the algorithm; it starts with infrastructure, data, skills and trust.”

His argument: Africa gained a billion people between 1980 and 2025, has 3,000 hours of sunshine a year to power compute infrastructure, and is building AI solutions for real problems — agriculture, health, financial inclusion. Smart Africa is also working on a “Compute Act” to harmonize compute centers across African countries.

A concrete signal from the forum: the on-stage signing of a memorandum of understanding between Afnic (France’s .fr registry) and Smart Africa, to network around twenty African national registries and organize South-South skills transfer on internet infrastructure. Digital sovereignty runs through these invisible building blocks too.

3. Training, the thread running through the whole day

If one word came up in every panel, it was skills. Three moments worth remembering.

“AI literacy needs to become as important as literacy itself.” The phrase is Henri Verdier’s, who added: “We don’t want to produce a humanity of passive consumers.” Knowing how to use AI critically is becoming a basic skill, on par with reading and writing.

Well-supervised AI helps learners progress. An edtech founder shared results from an experiment run with a major school: an AI-assisted learning program, supervised by a teacher, produced +25% success rate on assessments. But the same panel warned about the flip side: students who offload everything to the machine and lose their appetite for effort. The speakers’ message: the moment you hand over the ready-made answer, no learning gets built. Human support is what makes the difference.

Capacity building is becoming a global priority. An international network of AI capacity-building centers, backed by the UN, was announced at the forum. Every country should be able to train its own talent and produce its own models, rather than passively consuming those of others.

And for languages with little online presence — Wolof was cited as an example, but the same is true for Malagasy — the researchers on hand offered a simple truth: “The problem isn’t the language, it’s the entire culture behind the language — that, you can’t translate.” Without local production of content and knowledge, there’s no AI that looks like us.

What this means for you, SAYNA community

This forum confirms three convictions that guide our work every day:

Training yourself in digital and AI is no longer optional. When an ambassador compares AI literacy to literacy itself, and the UN launches capacity-building centers, the message is clear: digital skills are this decade’s ticket to entry.
African talent has a hand to play right now. Africa isn’t a spectator to the AI revolution: it’s becoming a player, with its infrastructure, its use cases and its talent. Every learner training today in Madagascar is part of this movement.
The human stays at the center. AI helps those who learn with it, guided and supported — not those who hand it their brain. That’s exactly the philosophy behind our paths: learn by doing, with real projects and real mentors.

The Francophonie wants to carry weight in global AI governance. It won’t do it without a generation of talent trained “on all five continents” — and that’s exactly our mission.

The full replay of the forum (8 hours of debates) is available on the OIF’s YouTube channel.

Want to catch the AI train instead of watching it go by? Check out our training paths at app.sayna.io.