Are African Teachers Ready for Artificial Intelligence? Decomposing Teacher Competencies and the Role of Initial and In-Service Training: Evidence from PASEC

Abstract

Enthusiasm for artificial intelligence (AI) in African education is outpacing the conditions required for AI to improve teaching and learning. Because large-scale African datasets do not yet measure AI use, this study reframes the question and argues that teachers' professional competencies are the prerequisite for any successful AI integration: AI does not replace the teacher but transforms the teacher's role, and its benefits depend on teachers' content and pedagogical knowledge, qualifications and continuing training. Drawing on the PASEC2019 teacher assessment across 14 Francophone African education systems (19,141 teachers), the study measures teacher content and pedagogical (didactic) knowledge, decomposes their variance across teachers, schools and countries using three-level models on plausible values, quantifies the contribution of pre-service qualification and in-service training with design-based estimates using PASEC replicate weights, and builds a Teacher AI-Readiness Index (TARI) to rank the systems. Results show that about 39% of the variance in teacher competence lies between countries, indicating highly unequal AI-readiness foundations. Holding a professional teaching diploma (associated with gains of +15 to +27 points across dimensions) and receiving in-service training (+9 to +12 points) are the strongest and most robust correlates of competence, together with classroom electricity, whereas the sheer duration of initial training or the number of in-service domains adds nothing. The ranking reveals wide gaps, from Gabon, Senegal and Benin at the top to Chad, DR Congo and Madagascar at the bottom. These findings align with, and can help target, the teacher-centred priority reforms that many countries are pursuing through their GPE partnership compacts. The study concludes that credible AI-in-education strategies in Africa must first invest in teacher qualification and continuing training, a condition, not a mere precursor, of AI’s effectiveness.