AI Unplugged: Bringing All Children Into the Future of Learning
As the world races toward an AI-powered future, in parts of the Global South, millions of children still attend schools without electricity or internet access. For them, the digital revolution feels distant, it’s invisible.
At this year’s Global Development Conference (GDC), which took place October 28–30 in Clermont-Ferrand, France, the 60 million girls Foundation led a hands-on workshop titled “Digital Offline Learning for Remote Communities – 10 Years of Field Experience Unplugged.” The session, organized in partnership with Brazil’s Center of Excellence in Social Technologies (NEES) and Change for Children, showcased how offline technologies and AI unplugged approaches can help make digital learning a reality for every child, regardless of where they live.
The Digital Divide That Persists
While the world celebrates breakthroughs in generative AI and adaptive learning, billions of learners remain disconnected. According to NEES, 20,000 schools in Brazil alone lack internet connectivity, and many more have bandwidth too weak for meaningful online learning.
In remote and rural communities from the Amazon to rural Africa, similar challenges persist: no electricity, no connectivity, few trained teachers. The result is a growing “AI gap” that mirrors, and deepens, existing educational inequalities.
As NEES researcher and UNESCO Chairholder on AIED uplugged, Ig Bittencourt explains, “We are basically replicating models that will make inequity even more resilient throughout the years.” When technology-driven policies focus only on connectivity and devices, they fail the most marginalized learners. Infrastructure investment alone cannot solve the digital divide—innovation must also adapt to the realities of low-resource contexts.
Offline Innovation: 60 Million Girls’ Mobile Learning Labs
Since 2014, 60 million girls, a Montreal-based organization dedicated to girls’ education, has been pioneering offline digital learning through its Mobile Learning Lab (MLL) model using a “plug and play” design, each MLL combines three core components:
- A RACHEL server that stores up to 1TB of curated educational content
- Up to 30 user devices—tablets, laptops, or smartphones—that connect to the server
- Solar panels to power the system in areas without electricity
These MLLs are deployed in over 20 countries, reaching more than 700,000 learners, mainly in remote or crisis-affected communities. The MLL enables access to rich, interactive learning materials in local languages, including government curriculum resources, ebooks, games, and even climate change modules. Evaluations show measurable gains in both cognitive and non-cognitive skills, with particularly striking results for girls: higher confidence, lower absenteeism, and greater comfort learning about sensitive topics like sexual health and menstrual hygiene.
As Wanda Bedard, President and Founder of 60 million girls, notes, “The social impacts of technology are at least as important as the academic ones.” When girls can access safe, relevant, and engaging content offline, they not only learn better, they see themselves as part of a digital future.
AI Unplugged: A New Vision for Inclusive AI
The concept of AI Unplugged aligns powerfully with 60 million girls’ mission. NEES has proposed five guiding principles for making AI accessible in resource-constrained environments.
- Conformity: Work with the infrastructure that exists—don’t wait for ideal conditions.
- Use the Internet only when available: Design systems that can function fully offline and synchronize later.
- Proxy users: When learners lack devices, empower teachers or local facilitators as intermediaries.
- Multi-user design: Move beyond one-to-one personalization to group-based learning models.
- Simplicity: Keep tools intuitive, since many teachers have limited digital literacy.
These principles redefine what AI innovation can look like in low-resource contexts. For example, NEES created an offline handwriting analysis tool that helps teachers assess student essays. Using only a smartphone camera and computer vision, it diagnoses writing levels and provides targeted feedback—no internet required. Similarly, lightweight AI models running on low-cost Android devices now assist teachers in generating math exercises or identifying students at risk of dropout.
Inclusivity Means Local Voices and Languages
As UNESCO’s recent report AI and the Future of Education: Confronting Coded Inequalities reminds us, inclusivity in AI is not only about access, it’s about representation.
When most AI systems are trained primarily on English or Western data, they risk excluding learners who think, speak, and learn differently. African and Indigenous languages, for example, often lack digital corpora, making them “invisible” to large language models. This creates a situation in which whole cultures are left out of how knowledge is digitally represented.
To avoid this, AI development must center local knowledge, multilingual design, and teacher collaboration. UNESCO calls for offline-compatible tools, human oversight, and participatory design—principles deeply reflected in both 60 million girls’ and NEES’s work.
A Shared Future: Equity by Design
At the GDC workshop on October 29, participants will have the opportunity to explore the Mobile Learning Lab, learn how AI Unplugged systems operate offline, and discuss what “inclusive digital transformation” truly means. Together, these initiatives offer a hopeful vision where AI and digital learning are not privileges of the connected few, but tools of empowerment for every child, everywhere.
Bridging the digital divide is a matter of justice. As AI reshapes economies and societies, inclusivity must be built into its foundation. The future of education depends on it.
Sources
Presentation: Bittencourt, I and Dermeval, D. (2025, Oct). AIED Unplugged: Expanding learning opportunities and reducing educational inequities in low- and middle-income countries.
UNESCO research: Nombuyiselo,V.M., Zondi, C., and Masikisik, B. (2025, Sept). Ensuring inclusive, contextualized AI in education: Considerations towards a roadmap. AI and the future of education: Disruptions, dilemmas and directions. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000395236
GDC abstract for Workshop 8 on Oct 29: Digital offline learning for remote communities – 10 years of field experience unplugged.