Côte d’Ivoire’s first AI strategy: Inclusive, safe and locally driven

As part of the Government AI Campus, Apolitical is publishing a series of articles featuring leaders from the 2026 Government AI 100, a list recognising public servants around the world who are pioneering AI adoption, capacity building and regulation.

Apolitical spoke with Rockya Coulibaly Fofana, a central figure in shaping Côte d’Ivoire’s approach to artificial intelligence and digital transformation, who is leading the nation’s first AI Strategy.

Can you tell us about the most important considerations that shaped your approach to Côte d’Ivoire’s AI strategy?

I think the most important consideration was to look at it as a national priority because, of course, we see everything that's going on in other countries, including countries that are more advanced. So we saw the need to bring in AI, but we wanted to bring it in an organised fashion. Not having people doing this and that, and with all the risks that that brings. So the idea was to think about our national priorities, put together a strategy in light of them, and consider the responsibility that, as a government, we have.

We also considered our implementation capacity and the government's responsibility to ensure that adopting AI wouldn't create more problems than it solves.

What was the most difficult trade-off you faced when developing the strategy?

I think that the most difficult trade-off was how do we manage speed because everybody's doing it, so we don't want to be left behind. But how do we also introduce safeguards without slowing down the adoption.

So the way we went about this is to help people adopt it, but at the same time develop the regulatory framework for this. So the ministry has been working with the national regulator so they can start drafting policies and rules for safeguarding while the ministry is pushing adoption.

How did you prioritise inclusivity when developing the roadmap for digital transformation and AI adoption?

Even before AI we had a digital gap here in at least two areas. First, the geographic gap, because in big cities you have an internet connection. In rural areas, you very rarely have a connection and the quality is not there. And then the other aspect is language. So in Côte d’Ivoire, we have about 60 local languages. The official language is French because of colonisation, so we go to school, we learn French, the government functions in French, but then more than half the population cannot use French as well as we wish.

So we're bringing in new AI tools and then suddenly more than half of the population is left out. So on top of the digital divide that we already had because of the connection, now we have another problem with language. So we see that inclusion has to be a pillar of strategy. That's why in the strategy that we developed, we had inclusion as a pillar. We have three pillars. One is infrastructure, second is investment and the third one is inclusion.

How do you solve these issues?

As a policy, we want inclusion. And in practice, what we're doing right now is that the ministry is supporting a private initiative that wants to create a local LLM for local languages. So the idea comes from an AI startup that's based here in Abidjan, but the government is fully supporting it.

The plan is to hire hundreds of people in the regions, and then they will start recording the names of things, the dictionaries, in local languages. It's probably going to take a few years to do it, but we started, that's the important thing there. And then once we have that kind of tool, the idea is to give access to other applications, other software that can come in through APIs and connect to it so that all the government services, for example, can be available to people who don't speak French, which is going to make a big difference in our country. We started with five languages, which are spoken the most.

Could you share an example of how AI has been used in one of those areas? What problem were you addressing, and how did AI help?

We've been working on an agriculture use case for more than a year now; it's an endeavour between the Ministry of AI and the Ministry of Agriculture. The idea is to use AI for crop yield prediction, climate risk modelling, and supply chain optimisation in the agricultural field.

Basically, we're trying to have as big a data set as possible. We're working with several international organisations. The World Bank is involved, the German Agency for International Cooperation (GIZ) is involved, and there's the United Nations Development Programme (UNDP). Many of the UN system organisations are helping us with this. So it's about collecting data.

What we've seen is that there's agricultural data going back to maybe the 1950s, but that was not used. It was just sitting in notebooks and such. So the first step was to digitise all these documents and then be able to use them for the purpose that we are trying to use it for. And so we're hoping that by the middle of 2026, the database can be made available to local startups that are developing solutions. So basically it's the government trying to put together the framework that companies can build upon to bring in solutions to the people in the rural areas.

What's one thing that every public servant should know or should learn about AI?

The important thing is safety. Even before we had a strategy, we could see that people were using AI tools. From the documents they were writing, I could see that this is just ChatGPT talking to me. That's not my co-worker talking. So people had just embraced the tool, and they're pretending that ChatGPT knows everything, and that everything ChatGPT says is true. So we have to break that. That was the first thing.

Besides the agriculture case, there's another initiative to drive adoption of generative AI tools in government, but we want to do it in an organised fashion. So we had a consultant create a portal for government employees. Through that portal, they can use the tools that they were already using in secret. Now they can use it in the open, and it's also set up so that the documents they upload don't leave. So in case somebody uploads a confidential document, at least it's within something that we have control over. The important thing is safety.

How do you expect AI use in government to change in two to three years? What can public servants start doing now to prepare for that?

We have quite a few training programmes for public servants so that they can be ready to see AI enter the sectors because right now I'm talking about the use of generative AI in general, but sooner or later the Ministry of Finance specifically will have its own AI tools. And then there's going to be the Ministry of Sports that's going to have its own AI tools tailored to their own sector needs.

So right now, we're just providing the basics so everybody can be ready to welcome the tools that are coming their way within the next maybe two, three, four years. Of course, there's going to be resistance. When I'm talking about the Ministry of Finance, of course more AI means more control, more surveillance in a certain way. Maybe not everybody's going to be happy about it, but they need to know that it's coming.

What excites you most about governments using AI and also what concerns you?

In terms of excitement, AI has the potential to enhance the quality of government services.

“I've seen the organised use of generative AI within our ministry. The documents that come out are much better. We can cut some timelines in half just because we're using AI in the right way. We usually talk about government being inefficient, but if we use AI the right way, we can become efficient, maybe as efficient as the private sector.”

But the risks cannot be underestimated. The wrong use and also the bias. Since most AI models are trained on data that comes from elsewhere, it's not adapted to the reality here. So there's always the risk that this bias could pollute the minds of younger people and perpetuate some bad behaviours that people already have. So we are indeed worried about that. And the hope is that raising awareness and putting together the right regulatory framework can alleviate this.