Embedded Files
How to use this page
How to use this page
If you have not seen our page on AI Education Equity in LMICs, we highly recommend reviewing it first, as it provides important background before exploring these examples.
This page is not a central resource hub but rather a place to showcase examples of equitable AI tools and methods that support students and teachers in their own contexts. These examples highlight how developers and educators have taken into account both the data fed into the systems and the technical constraints such as using platforms like WhatsApp or designing with an offline first approach.
These tools and methods will continue to evolve, but as long as you remain mindful of the data you provide and consider the technical constraints of your environment, you will be in a strong position to make good use of them.
Teacher-Focused AI Tools
Teacher-Focused AI Tools
(lesson planning, assessment, PD, classroom support)
TheTeacher.AI (Sierra Leone)
TheTeacher.AI (Sierra Leone)
A WhatsApp‑based chatbot that provides on‑demand professional development. It generates lesson ideas, answer explanations and classroom management tips. Designed to work under poor connectivity (WhatsApp was chosen because it performs well on intermittent networks and uses little data); teachers already own phones, so no extra hardware is needed (gaied.org)
Shiksha Copilot (India)
Shiksha Copilot (India)
Co‑developed by Microsoft Research and Indian teachers, this AI assistant helps create lesson plans, assignments and hands‑on activities. It ingests textbooks and local syllabi and can deliver content over WhatsApp and Telegram. During pilot testing in Bengaluru schools, teachers reported that planning time dropped from over an hour to 60–90 seconds (itbrief.co.nz)
Camara AI Assistant (East Africa)
Camara AI Assistant (East Africa)
Locally integrated, fast, offline AI solution designed for both teachers and students. Generates lesson plans, quizzes and answers aligned to national curricula; supports literacy, coding and creative learning; suggests follow‑up topics; and integrates local textbooks while ensuring privacy because it runs entirely without internet. Camara aims to provide schools in Kenya, Ethiopia, Tanzania and Zambia with secure AI‑assisted tools tailored to local needs
SwiftPAL (India) – ConveGenius
SwiftPAL (India) – ConveGenius
Data‑driven “internet‑free” tablet classroom. Provides personalised lessons during school hours and gives teachers actionable, student‑level data to support multi‑grade classrooms. SwiftPAL has reached over half a million students across 15 Indian states and is designed to break cycles of poverty.
21st Century Teacher Educator GPT (Ghana)
21st Century Teacher Educator GPT (Ghana)
Custom GPT built using Retrieval‑Augmented Generation. It embeds Ghana’s national teacher‑education curriculum, UNESCO AI guidelines and culturally responsive pedagogies. Pre‑service teachers enter their institution, year and semester; the tool produces lecture notes, assessment practice and research guidance in English or Ghanaian languages like Twi, Dagbani, Mampruli and Dagaare (arxiv.org)
Student-Focused AI Tools
Student-Focused AI Tools
(adaptive learning, tutoring, accessibility, engagement)
Rori (Ghana)
Rori (Ghana)
WhatsApp‑based AI math tutor from Rising Academies. Provides micro‑lessons, interactive exercises and socio‑emotional coaching. Runs on low‑bandwidth networks and basic phones; has served over 100,000 students and emphasizes growth mindset (the-learning-agency.com)
RoboTutor (Tanzania)
RoboTutor (Tanzania)
XPRIZE‑winning Android tablet app that helps children learn reading and numeracy without adult assistance. Uses language technologies and machine learning to adapt to each child (xprize.org)
onecourse / Onebillion (Malawi)
onecourse / Onebillion (Malawi)
Adaptive literacy and numeracy programme available on rugged, low‑cost Onetab tablets. Runs fully offline, adjusts to each child’s level and is available in multiple languages
Kolibri (Global) – Learning Equality
Kolibri (Global) – Learning Equality
Offline‑first platform providing a digital library and adaptive quizzes. Students connect to a classroom server via tablets or laptops, work through lessons at their own pace and receive immediate feedback while teachers track progress
M‑Shule (Kenya)
M‑Shule (Kenya)
SMS and WhatsApp platform delivering micro‑courses, assessments and personalised tutoring in multiple languages (English, Kiswahili, Dholuo, Kamba, Kikuyu, Ng’aturkana, Somali). It targets households without smartphones or internet and has reached 13,000 households across 30 Kenyan counties
Kwame for Science (West Africa)
Kwame for Science (West Africa)
Sentence‑BERT‑based web app answering science questions. It presents relevant paragraphs and top past exam questions and achieved 87.5 % top‑3 accuracy during a two‑week pilot across 11 African countries (ceur-ws.org)
Ferby Chatbot (India)
Ferby Chatbot (India)
Multilingual AI tutor deployed by Education Above All, providing context‑sensitive help for millions of Indian students. Limited information is publicly available, but reports note that it supports local languages and aligns with national curricula.
Access pathways
Access pathways
Access pathways refer to the strategies and mechanisms that enable communities in low- and middle-income countries (LMICs) to use and benefit from AI tools equitably. This includes lowering barriers such as connectivity, affordability, and availability of localized resources.
WhatsApp
WhatsApp
Widely used and often zero‑rated in LMICs; it serves as the delivery channel for TheTeacher.AI, Rori and Shiksha Copilot, allowing teachers and students to access AI using existing data plans
Telegram and SMS
Telegram and SMS
Deliver lessons when WhatsApp is unavailable; M‑Shule uses SMS for micro‑courses
Access Offline AI Model
Access Offline AI Model
Offline models are AI systems designed to function in low-bandwidth or no-internet environments, giving users the ability to run tools locally. They support local ownership and control of models
PocketPal AI
PocketPal AI
Open‑source app that runs small language models (SLMs) locally on smartphones without internet. Users can download models (Phi, Gemma, Qwen etc.), create AI “personas” for tasks and adjust inference settings; all conversations stay on the device
LM Studio
LM Studio
Desktop application allowing users to download and run large‑language models (e.g., GPT‑oss, Qwen, Gemma, DeepSeek) on local machines; free for home and work use
Whisper.cpp
Whisper.cpp
lightweight C/C++ implementation of OpenAI’s Whisper speech‑recognition model. Runs fully offline on devices like iPhones and can be integrated into apps, enabling transcription and voice interfaces in classrooms
Page updated
Google Sites
Report abuse