Intelligent Tutoring, Adaptive Learning and Educational Equity: A Conceptual Model of Personalisation, Access and the Risk of a Two-Tier Education System

Abstract

Adaptive learning systems and intelligent tutors are marketed as a tutor for every child, a promise most persuasive where teachers are scarce, classes are large and foundational learning is weak, as in much of sub-Saharan Africa. Their equity consequences, however, have been theorised far less than their average effects. This conceptual paper addresses that gap. Its objective is to explain when adaptive learning narrows or widens learning gaps and to derive testable propositions and deployment principles. The population is the peer-reviewed literature on intelligent tutoring and adaptive learning, educational equity and the digital divide, and educational technology in low- and middle-income countries. The sample comprises 35 sources: 33 journal articles published 2023–2026, including trials and quasi-experimental and implementation studies from Kenya, Sierra Leone and Tanzania and longitudinal evidence on remote learning in Ghana, and two foundational works. Sources were chosen by criterion-based purposive selection, each verified against its Crossref record and screened for retractions; preprints, reports and conference papers were excluded. Analysis was a conceptual synthesis: empirical sources were coded for what was personalised, teacher involvement, setting of use and the distribution of outcomes, and propositions were derived abductively. The paper distinguishes personalisation of pace, content and pedagogy, and argues that current systems supply the first readily, the second partially and the third rarely, so that the learners who most need responsive teaching receive least of it from software. A three-tier access model then makes the equity effect depend on who gets the technology, who gets the technology plus a teacher, and who gets neither. Seven falsifiable propositions specify when adaptive systems narrow or widen gaps, including the risk that software becomes the education of the poor while advantaged learners keep human teachers. For low-resource, especially African, school systems the paper recommends teacher-complementary design, protected in-school dosage, language and curriculum alignment, low-technology channels and distribution-sensitive monitoring, and it sets a research agenda on the teacher's role, learner motivation and the distribution of outcomes.