An Artificial Intelligence in Education Community of Interest Event
As AI tools become an increasingly visible presence in schools, the question of how to prepare teachers to engage with them critically, confidently and safely has moved to the center of education policy debates. But building genuine AI literacy goes beyond training teachers to use tools designed without them. It requires putting teachers at the heart of developing AI systems and resources that shape their classrooms. Our latest AI in Education Community of Interest (CoI) session, co-organised with the World Bank and EdTech Hub’s AI Observatory & Action Lab, brought together experts from Central Square Foundation, OpenAI and Rising Academies to explore: What does meaningful AI literacy look like for teachers?
The conversation spanned the panellists’ experiences in India, Rwanda, Estonia, Uruguay, and beyond. The speakers explored what genuine AI literacy looks like for teachers—from foundational concepts to classroom practice—and what it takes to build AI literacy, agency, and AI tools themselves in ways that are grounded in teacher knowledge, not imposed from above.
Featured Speakers
- Opening – Mike Trucano – Senior Advisor for Education, Technology & Innovation, World Bank
- Moderator – Flic Burgess, Teachers-in-the-Lead Sandbox Lead, EdTech Hub – AI Observatory and Action Lab
- Fidele Hagenimana, Head of Programs, Rising Academies
- Gouri Gupta, Senior Project Director, EdTech and AI, Central Square Foundation
- Jayna Devani, Lead, Education for Countries, OpenAI
Watch the Webinar
Key Takeaways
Here are four key takeaways from the discussion on AI literacy for teachers which illustrate the gap between usage and genuine literacy, the case for teachers as co-designers rather than merely end users, the challenge of keeping pace with a rapidly changing landscape, and what low-resource settings are teaching us about what actually works.
1. High usage, low literacy – the difference matters enormously
One of the most striking data points of the session came from Gouri Gupta at Central Square Foundation (CSF), whose team conducted a large-scale survey of teacher AI usage in India. The findings were revealing. For the first time, technology usage was higher among teachers than students. But when those same teachers were asked how AI works, roughly half described it as a search engine.
The gap between usage and literacy is not a minor distinction. A teacher who uses AI tools every day can still be accepting flawed outputs, missing bias, or failing to notice where the tool breaks down because they lack the conceptual grounding to evaluate what they are seeing. This insight drove CSF to develop AI Samarth, a contextual AI literacy curriculum now reaching 2.2 million teachers across India. It is anchored in a practical definition: using AI responsibly, ethically, and meaningfully, not just skilfully.
Gouri outlined the pillars that emerged from that work: understanding what AI is; grasping its practical and ethical implications; using it to make work more efficient and meaningful; and, critically, knowing how to guide students in using it well. That last element was the one teachers felt least equipped to address. They could see students using AI but had no framework for how to support them. Closing that gap is at the core of what genuine teacher AI literacy requires.
2. Teachers are not only end users, they are key stakeholders and co-designers
A consistent thread across the session was a sharp critique of the dominant model of AI tool deployment: tools designed without teachers, then handed to them for adoption. Fidele Hagenimana from Rising Academies offered a vivid alternative from their work in Rwanda with TARI, a WhatsApp-based teacher support tool built with teachers as co-designers from the ground up.
The result was a process full of productive friction. Teachers pushed back on lesson structures that did not reflect their classroom realities; they wanted the tool to respect their professional judgment, not replace it. In response, TARI was redesigned to ask contextual questions before generating content. For instance, how many students are in the class, and are any students living with disabilities?
Teachers also flagged that the pronunciation guidance reflected a Western English accent that did not match their context in Rwanda, where the language of instruction recently shifted to English. As a result, the tool was updated to reflect East African speech patterns.
The principle Fidele drew from this experience is one he argued should guide the field at scale: teachers must be at the centre of AI development if tools are to genuinely strengthen education systems. What scales is not the specific solution, but the participatory spirit and the continuous local feedback loops, with consistent input from educators.
3. Teachers need to have a holistic understanding of AI and its affordances to avoid needing to re-learn new tools
One of the sharpest exchanges of the session centred on a genuine structural challenge. AI tools evolve by the week, while teacher training programmes take months to design and roll out.
Jayna Devani from OpenAI acknowledged this directly. Governments across the nine countries where OpenAI has formally deployed ChatGPT in schools are grappling with it in real time. They know teachers are already using ChatGPT, Claude, NotebookLM, and other tools informally every day, and they are asking what their institutional role should be.
Her answer was to resist the temptation of tool-specific training. The goal instead is to build teacher agency and confidence to judge what is useful, what is safe, and what is pedagogically sound—capabilities that will remain valuable as technology continues to change.
4. AI is reaching low-resource settings, but sustaining engagement among teachers is a challenge.
The session closed on a practical question: what works for teachers in low-connectivity settings, and what is still difficult?
Fidele’s answer from the field was direct. Tools like TARI, running on WhatsApp with basic phones, are already reaching teachers in rural contexts. Simplicity—through short lesson prompts, micro-assessments, and quick feedback that fits into existing routines without requiring expensive devices or constant connectivity—is what makes it work. Teachers adopt these tools when they fit into daily life, not when they add to an already full workload.
What remains hard, Fidele argued, is sustaining engagement beyond the initial novelty and building the kind of trust that makes ongoing use possible. Teachers need to know their data is safe and that the tool respects rather than diminishes their professional role.
Jayna offered a complementary perspective. In low-resource environments, the most meaningful AI contribution may not always be a chatbot in front of every teacher. She pointed to a pilot with Uruguay and UNICEF exploring whether AI can accelerate the production of accessible digital textbooks—a process currently taking six to nine months and tens of thousands of dollars per book. If AI can collapse that cost and timeline, the system-level impact could be significant.
The deeper lesson is that AI’s value in low-resource settings depends on situating it within a broader conversation about what the whole system needs, not just what the tool can do.
Further questions from the session
The following questions were posed by Community members. We are sharing them to help stimulate further discussion and knowledge exchange. Please note that some questions may have been edited for spelling or clarity.
Teacher agency, responsibility, and the risk of over-reliance
- How can teachers use AI effectively and ethically without becoming AI operators, performing the labour of policing and correcting outputs on behalf of AI companies while students disengage from genuine thinking and learning?
- Teachers are often pitched AI as handling the tedious work—lesson planning, feedback, and assessment design—so they can spend more time with students. Those are, however, the tasks through which teachers develop their craft. What does “more time with students” look like in practice, and where does the teacher’s role go in five years?
- How do we build in system-level responsibility for teacher capacity so that the burden of keeping up with AI developments does not fall unfairly on already overstretched teachers?
Inclusion and disability
- What kinds of training, tools, or system-level support have helped teachers build their own AI literacy while also strengthening support for students with disabilities?
Evidence
- What would a genuine tipping point look like? What evidence would be most convincing that AI is helping teachers, particularly in LMIC contexts?
Opportunity
EdTech Hub’s AI Observatory & Action Lab is hosting an immersive online Field Trip to the Future that takes you to 2031, a world where AI handles most of the work and invites you to engage with that future as if it is already here.
By the end of the session, you will have worked with others to identify a shared set of skills that are grounded in a possible future, not just current guesswork. You will also leave with a question worth carrying back into your own work: what would need to shift today to make sure every learner has access to them?
Please note, because this is an interactive session we only have a maximum of 80 available slots. Registration is first come, first served, so leave your AI assistant at home and come ready to travel to the future!
2–3:30 PM BST | 18 June 2026 | Online
Resources
The following resources were shared by Community members and participants. These have not been reviewed by the World Bank or EdTech Hub, but are useful indicators of what conversations, evidence, and methods are being explored in the sector.
Resources from Rising Academies
- TARI — Teacher AI tool for Rwanda classrooms
- RORI — AI tool for students
- LearnLens — AI-powered assessment tool for teachers
- Partnering for Progress: AI and the Future of Learning in Rwanda
Resources from Central Square Foundation
- AI Samarth — AI Literacy Curriculum for students and teachers
- AI Samarth Curriculum Framework
- CSF EdTech and AI Programme
- Casebook on Real-World Impact of AI in Education (March 2026)
- EdTech Tulna — Quality standards for EdTech
Resources from OpenAI
- Education for Countries programme
- OpenAI Academy — K-12 Education Community
- Understanding AI and Learning Outcomes
- ChatGPT Foundations for Teachers (Coursera)
Other resources
Resources from the World Bank
- How to use ChatGPT to support teachers: The good, the bad, and the ugly
- Teachers are leading an AI revolution in Korean classrooms
Resources from EdTech Hub
This is part of an ongoing series hosted by the World Bank and EdTech Hub’s AI Observatory and Action Lab. The AI Observatory is made possible by support from UK International Development. Please follow along and join the conversation on LinkedIn!
