Schools Don't Need Better AI Technology. They Need Better Organization
The infrastructure gap is not why most education systems can't use artificial intelligence effectively. The real problem is organizational culture.
Education leaders often frame AI as a classroom problem: devices, chatbots, adaptive platforms. This framing creates a dead end. Schools in low- and middle-income countries lack the infrastructure for these solutions, so AI becomes something only well-resourced systems can afford. Everyone else waits.
But a different approach exists. World Bank research distinguishes between "big AI"-massive compute power and specialized talent-and "small AI," practical task-specific tools that run on everyday devices. Small AI is already transforming agriculture and healthcare in developing countries. In education, it could do the same without putting new hardware in classrooms.
What Small AI Actually Does in Schools
Consider a concrete example from a Latin American country. Teachers need to know not just which students are falling behind, but what misconceptions they hold. This requires diagnostic questions where each wrong answer reveals a specific gap. Most systems don't have them because producing thousands of aligned items is slow and expensive.
Four teachers wrote seed questions based on real classroom errors. They used Claude Code-a tool that lets non-developers give AI instructions in plain language-to read a 225-page national textbook and scale those seeds into 3,950 diagnostic questions. Every step included human review: experts chose which errors mattered, reviewed what the AI produced, and rejected what didn't work.
The project took six weeks and four people. The alternative wasn't a handcrafted process by a large team. It was teachers starting the year without knowing where each student stood, relying on intuition.
Small AI didn't replace an expensive process. It replaced the absence of one.
The Imagination Gap
Most education institutions don't know these applications are possible. IT departments and curriculum teams sit in different meetings. Nobody has shown small AI to the people who actually run education processes.
The untapped opportunities are straightforward:
- Regulatory compliance. Countries accumulate thousands of overlapping rules across government levels. Feed those regulations to Claude Code and it flags contradictions, identifies duplicates, and produces a simplified guide in days.
- Workflow mapping. Ministry procedure manuals can be analyzed to identify bottlenecks and suggest streamlined alternatives. Not replacing judgment, but showing where time is wasted.
- Differentiated training. Most ministries deliver one training to all teachers because designing versions for different schools takes more time than any team has. Claude Code can adapt core content for each school profile, adjusting examples and focus areas.
- Principal reporting. Principals spend 76 percent of their time on paperwork. What if they sent a voice note about the week and AI transcribed it into the compliance report the ministry requires? Better data, because principals report honestly instead of copying last month's template.
What Organizations Actually Need
Most ministries lack three things that organizations need to capture AI gains: leadership willing to experiment, a lab that turns ideas into solutions fast, and staff using these tools daily.
A starting point is modest. Pair one person who is AI proficient-not a software engineer, but someone comfortable enough with Claude Code to turn a real problem into a working prototype-with people who know the problem deeply. A curriculum specialist. A training coordinator. A supervision lead.
Pick one task: turning 200 school visit reports into a regional summary that identifies which schools need support and why. Run it with Claude Code alongside the staff who normally do it. Compare results. AI suggests, humans evaluate.
Any pilot must follow proper data privacy protocols. Starting small is how you manage that risk. If results hold up, invest further. If not, the cost was weeks, not years.
The Real Constraint
The grand vision of AI-powered classrooms may arrive gradually, grounded in evidence about what actually improves learning. That will require full connectivity, devices, cheaper data processing, and deep AI literacy among teachers and students.
Small AI requires none of that from schools. It does require AI-proficient people inside the institutions that support them. People comfortable enough with the tools to turn a real education problem into a working prototype. People working with cross-disciplinary teams that know the subject, the students, and the teachers involved.
The binding constraint is not technology. It is whether institutions can create the conditions for those people to work together with the right support.
For education leaders interested in building this capability, Claude AI Courses and AI for Education resources offer practical starting points for developing the skills your team needs.
Your membership also unlocks:
