Teaching Through War With AI
Schools closing during air raids. Teachers scattered across borders. Electricity cutting out mid-lesson. That's daily life for educators in Ukraine, now in its fourth calendar year of full-scale war with Russia.
In this context, one question matters: can AI help sustain teaching and learning under crisis conditions? At the Harvard Graduate School of Education's January conference led by Ford foundation professor Fernando Reimers, a student panel dug in, focusing on practical ways AI can support teachers and students amid instability.
What's happening on the ground
Since 2022, more than 43,000 Ukrainian teachers have been internally displaced or forced to leave the country. Nearly a third of classes now run fully online or in hybrid formats. Schools have been damaged or destroyed, power fails often, and connectivity is spotty-especially near front lines and in rural areas.
Digital tools, including AI, were once optional. Today, they function as essential infrastructure to bridge teacher shortages and keep instruction moving.
Who's stepping in-and the limits
Global tech companies and international organizations have provided devices, infrastructure, and training. In 2022, Google and UNESCO supplied 50,000 computers to support distance learning. More recently, AI companies announced programs to spread AI literacy and tools to educators worldwide.
But investment doesn't automatically translate into impact. As HGSE student Chloe Zeng put it, "The problem isn't a lack of pilot programs or innovation. It's a lack of coherence." Without a clear national framework, strong training institutions, and equitable distribution, well-resourced regions move forward while vulnerable communities stall.
What's missing right now
- No shared roadmap: There's no national agreement on teacher AI competencies, assessment, or how initiatives fit together.
- Unstable infrastructure: Power cuts and destroyed facilities break continuity, making online-first plans unreliable.
- Inequitable access: Teachers with better connectivity and mobility gain from AI; those in conflict zones are left behind.
Policy options the team evaluated
- Train 10,000 "AI Master Teachers" via a voluntary micro-credential program.
- Stand up 35 regional AI innovation hubs offering intensive, in-person coaching.
- Launch a centrally guided national AI upskilling program delivered through existing institutions.
The recommendation: lead with the national upskilling program for coherence and reach, then layer targeted, high-intensity support where conditions allow.
What education leaders can do now
- Define baseline AI competencies for teachers (lesson planning assistance, formative feedback, offline prompts) and map them to subjects and grade bands - see the AI Learning Path for Primary School Teachers for a practical example.
- Adopt offline-first and low-bandwidth tools. Prepare printed packets and SMS/USSD backups for instruction during outages.
- Create simple continuity plans: lesson "playbooks," power/backup tiers, and phone trees for teacher coordination.
- Centralize trusted resources. Curate a lightweight repository of model prompts, rubrics, and vetted AI use cases aligned to curriculum.
- Use train-the-trainer models through existing institutions. Build regional cohorts that can deliver micro-workshops in local languages.
- Measure what matters. Track student attendance, short-cycle learning checks, and teacher workload changes-not just tool adoption.
- Prioritize equity. Allocate devices, connectivity support, and coaching to frontline and rural schools first.
- Protect students and staff. Set clear data, safety, and content guidelines; require human oversight for grading and sensitive decisions.
- Support teacher wellbeing. Keep AI use focused on reducing admin load, generating lesson starters, and offering feedback-freeing time for human connection.
Why this matters beyond Ukraine
AI is now a core policy issue across income levels. At HGSE, student work also examined under-financed classrooms in Uruguay, conflict-prone provinces in Thailand, and remote learning in Papua New Guinea. The throughline: technology can keep learning afloat, but only when paired with coherent policy, stable delivery channels, and focused teacher support.
AI isn't a cure-all. It's a practical set of tools. The job is to make those tools reliable, equitable, and simple enough to use under pressure.
For teams building capacity
If you're planning professional learning around AI-by subject, skill, or role-this curated directory can help you scan options and structure pathways: AI courses by job. For technical teams designing or evaluating tools, the AI Learning Path for Research & Development Engineers offers a complementary pathway for building developer and R&D capacity.
Your membership also unlocks:
