With the development of data acquisition technologies such as the Internet of Things, digital twins, and others, large volumes of data are produced daily, and leveraging them can significantly aid decision-making. Many applications of these analyses are taking place in the African context: artificial intelligence (AI) tools can help companies exploit the data accumulated on their servers.

The University Clermont Auvergne (through its LIMOS and ISIMA components) and the University of Yaoundé I (through its computer science department and its engineering school Polytechnique of Yaoundé ENSPY), in partnership with the Cameroonian AI scholarly society (CAIS), organized in 2024 a thematic school on artificial intelligence I3AFD (https://i3afd.sciencesconf.org/) over the course of a week, with the in-person participation of 35 Master's and PhD students, and 150 remote learners. This school was part of the FDMI-AMG research project funded by the CNRS’s DSCA program.

The successful organization of this school has encouraged us to plan a new edition, which we envision for Fall 2025. A discussion with the members of the scientific committee of the I3AFD 2024 school, as well as with the learners, led us to focus this second edition on the theme of generative AI, with an emphasis on under-resourced languages, which are a key characteristic of the languages in African countries, particularly those in Central Africa.

Generative AI and large language models, including ChatGPT, have ushered in a new era with rich and diverse capabilities for numerous application domains. It is therefore essential to introduce the foundations of large language models (LLMs), to discuss the infrastructure, tools, and best practices for designing and running applications using LLMs, and to explore various techniques to use and build specialized models for specific tasks and/or with limited data (small language models - SLMs).