“AI has the most benefit in the complex environment of developing economies”

How would you define artificial intelligence (AI)?

Maha Jouini: With AI, machines are given a form of consciousness and start making decisions in place of human beings. They are able to take action.

Seydina Ndiaye: AI is a capacity given to computers to have a behaviour that resembles human behaviour.

What opportunities does AI offer for developing economies?

Seydina Ndiaye: The deployment of AI is not facilitated by the complex environment of developing economies. Indeed, AI requires high levels of computing power, energy, and investment which these countries don’t have. But AI has the most benefit for them: it is diverse and offers a wide range of usable tools. Certain technologies require fewer resources than others, while using a fairly advanced AI, which has already been tried and tested elsewhere and can be geared to address the specific problems of developing economies. For example, in the field of health, the Senegalese start-up KERA has made it possible to reduce the face-to-face time between the patient and their doctor by collecting a range of information beforehand, which facilitates an initial diagnosis. In the agriculture sector, we have a project in Senegal that maps rice crops and estimates their yields.

However, in these countries where AI is a new subject, we need to be able to set up the necessary systems in terms of human resources able to develop and deploy these solutions. A sufficient computing and storage capacity is also required, as well as at least a minimum of supervision to ensure that AI technologies are used safely and ethically. We set up models that need to be tested before being deployed among the final users. This requires the establishment of regulations at the highest level, the State, to control the deployment of certain tools.

How can AI be a driver for inclusion for women and young people?

Maha Jouini: AI is currently not a driver for inclusion, it is the mirror of an unfair reality, made up of stereotypes and prejudices. It firstly defines me as a woman, a Muslim, and an Arab, and these three identities have sometimes hindered my career: AI will not be a driver until it reflects the full complexity of a human mind.

In Africa, women don’t have access to higher education, they don’t have access to Internet, and they often live in vulnerable conditions. The digitalisation underway is not at all equitable. According to UNESCO, women represent less than 20% of AI experts. AI will be able to become a driver for women and young people when it has been provided with good governance. It’s through digital sovereignty that underdeveloped countries will be able to emerge from poverty, create development opportunities, and minimise their debt. Without this sovereignty, we’ll experience a digital colonisation and AI will increase the control that big data companies have over our economy.

How can AI still be used to achieve the SDGs?

Seydina Ndiaye: AI can have a significant leverage effect. It can help to accelerate certain actions, for example, in the fields of education and food for all. But it cannot singlehandedly provide all the solutions for the achievement of the SDGs. There are other factors involved, such as public management and the economy. In addition, if we map the projects related to AI and the UN’s SDGs, most are confined to three sectors, agriculture, health and education, because they are the most lucrative. Companies and researchers focus on these sectors, whereas the challenge facing everyone lies in ensuring that research is also developed in other sectors for the benefit of people.

What ethical principles should guide the development and application of AI?

Maha Jouini: As I said before, without good governance, AI can be a harmful tool for humanity. In Africa and in Arab countries, where the digital transformation is not yet complete, AI requires an institutional framework, national strategies and policies, and legislation. For example, there are still not enough laws on cybercrime and data security. We’re pushing for the development of communication systems and access to Internet in the most remote regions, while the AI era has already begun: we’re in a race against time! Up until now, the digital transformation has passed through external experts, but the problems related to digitalisation differ depending on the context: each country needs to be able to develop its own solutions to integrate AI, and implement its own coding and regulation systems. We need more training for this.

What are the main technical challenges in the development of ethical and inclusive AI technologies?

Seydina Niaye: Most of the technical challenges concern infrastructure, telecommunications and energies, for example, which remain very expensive in emerging countries, whereas they are essential for adapting what already exists and making it usable. Most data are also expensive to acquire, store and process, which creates a certain disparity in access to this data. We also need to remain vigilant throughout the value chain, from the acquisition of data to their labelling: they can also be misused if they aren’t controlled.

Finally, inclusive AI is part of the SDGs at the highest level of the UN: everyone must be able to benefit from it. The biggest challenge today lies in making developing countries aware and getting them to take the plunge: it is important for these countries to take an interest in AI and take advantage of the opportunities it offers.

How can we ensure that AI systems are understandable for non-specialists?

Maha Jouini: We do indeed need to democratise AI and to do so, Arabs and Africans cannot use Western narratives as a basis: they need to create their own language, and their own definitions of AI and its governance. The digital transformation of AI needs to be carried out from the ground up. It needs to be cross-cutting and multi-dimensional. AI is not just a matter for specialists, it also needs philosophers, sociologists, legal officials, and even clerics. We’re in the era of the fourth industrial revolution, AI is everyone’s business. I really encourage Africans and Arabs to take up AI: it represents the era of intelligence, the era of “minds”, science has no colour, gender or religion. We’re all intelligent, we can all use AI for our development and we mustn’t underestimate ourselves.