This case study examines AI Education Unplugged, an approach developed and advanced by Dr. Seiji Isotani, Faculty Director of the AI and Learning Analytics Program at the University of Pennsylvania Graduate School of Education and President of the International Artificial Intelligence in Education Society. The case study is originally based on an interview with Dr. Isotani conducted by the author, and is further informed by his presentation materials and the published research associated with the project.

The central question is straightforward but globally urgent: how can the benefits of AI in education reach people and places where technology and infrastructure are absent, unstable, or severely limited? Rather than assuming one-device-per-student classrooms, reliable broadband, or advanced digital skills, AI Education Unplugged starts from the infrastructure that already exists in underserved settings. Its design logic emphasizes available infrastructure, intermittent connectivity, shared resources, and proxy use through teachers or other intermediaries.

The case is grounded in the Brazilian context, particularly in the National Policy on Learning Recovery following the pandemic. In this setting, Dr. Isotani and collaborators applied the AIED Unplugged approach to support student writing development at scale. Instead of relying on fully digital classroom systems, students completed writing tasks on paper, teachers or school staff captured those responses through mobile devices, uploads occurred whenever internet access was available, and AI-supported analysis returned diagnostic feedback and teacher-facing dashboards. This substantially reduced delays compared with older review systems and better aligned educational support with real classroom conditions.

The case highlights a broader policy argument: mainstream AI in education can become a new source of educational inequality if it assumes infrastructure that many communities do not have. AI Education Unplugged offers an alternative path by redesigning educational AI for low-resource environments. The case also draws on cited published research shown in Dr. Isotani’s slides, including work on educational inequality, mobile-phone access, delayed internet convergence, writing-support systems in Brazil, and teacher-mediated AI use in underserved settings.

This case study is presented as an original analytical synthesis, not a reproduction of Dr. Isotani’s slides or publications. All data, claims, and conceptual frameworks should be attributed to the underlying published sources and to the interview where relevant, in order to preserve academic integrity.