STEPS — AI-Assisted Science Curriculum Cuts Development Costs 70% for Benin's Primary Schools

Launched in 2021 with support from the Global Partnership for Education's Knowledge and Innovation Exchange (GPE-KIX, managed by Canada's International Development Research Centre), the STEPS project set out to close a persistent gap in primary science teaching materials across Benin, Cameroon and the Democratic Republic of Congo. Implemented by the Frank Foundation and NextGenU.org together with national partners, the project combined generative-AI drafting and translation with mandatory human review by curriculum specialists to produce culturally relevant, curriculum-aligned science textbooks and teacher guides.

The AI-assisted workflow — drafting initial content against national curriculum requirements, proposing locally relevant examples, and supporting translation and localisation — reduced content development time and cost by up to 70% compared with conventional production methods, according to the implementing partners. The project produced 28 science textbooks and workbooks plus 52 culturally adapted workbooks and teacher guides, translated into French and, in Benin, the local Fon language.

By the end of the initial phase the materials had reached over 6,000 students and 160 teachers across 18 schools. In Benin specifically, the Ministry of Education initiated formal procedures to authorise the STEPS science textbooks for national use — a stronger uptake signal than the pilot-only status typical of many donor-funded EdTech projects. Partners in Cameroon and DR Congo have taken similar or further steps, with the DRC partner expanding the material to all 101 of its schools.

Published figures on cost and time savings come from the implementing partners rather than an independent audit, and no student learning-outcome data (test scores, retention) has yet been published — the evidence to date documents production efficiency and institutional uptake, not classroom learning impact.