As someone who works closely with teachers and teacher professional development, I cannot help but wonder: Could artificial intelligence (AI) help teachers find new routes for their professional learning? Could it offer direction when conventional opportunities are difficult to reach? Or might it create new risks and inequalities?

These questions took me back to my childhood. When I was in grade four, one of my teachers suddenly stopped coming to the school from home for almost 10 months. I had no clue about his whereabouts. When he finally returned, I asked him where he had been. “You come to school to learn,” he said. “I also had to go to my school to learn.” His answer surprised me. Until then, I had believed that teachers knew everything. I had never thought that they, too, needed to learn, practise and update their knowledge. That brief conversation showed me something I have never forgotten: good teachers never stop learning.

Years later, as a teacher trainer, I worked with more than 5,000 English teachers across Nepal, including teachers in remote mountain communities, rural municipalities and rapidly growing cities. Many were deeply committed to their students. Yet they often had limited access to relevant training, professional resources, mentoring and follow-up support. One question stays with me: How can professional learning become more accessible, relevant and responsive to teachers’ needs? Technology gradually became part of the answer. Digital tools have allowed teachers to find resources, connect with colleagues and explore new classroom practices. 

Now, AI is taking these possibilities much further. Within seconds, a teacher can generate a lesson plan, simplify a difficult concept, translate terminology, create classroom activities or receive feedback on a piece of writing. This sounds exciting. But is it the full story?

When I began using AI for my own professional learning, I felt both eager and frustrated. It helped me find information, explore ideas and improve materials. However, it also produced inaccurate, culturally inappropriate or simply unhelpful answers. I soon realised that using AI successfully involves much more than typing a clever prompt. It requires professional judgement, ethical awareness and careful verification.

Most importantly, it requires understanding the learners, classrooms and communities for whom the materials are intended. I was not alone. Teachers were already experimenting with AI, often quietly and without formal guidance. But how widespread was its use? What were teachers using it for? Who was benefiting? Who was being left behind? And was Nepal’s education system ready to support its responsible and fair use?

A recent joint study by the Ministry of Education and Sports and the British Council explored these questions. Its data are both encouraging and challenging. The study surveyed 815 teachers across all seven provinces. It combined this national picture with interviews, focus group discussions, classroom observations, stakeholder roundtables and action research led by 21 teachers.

One finding stood out: 76 percent of the surveyed teachers had already used generative AI. The central question, therefore, is no longer whether AI will enter Nepal’s classrooms. It is already there. Teachers reported using tools such as ChatGPT, Gemini, Google Lens, Google Translate and others to find information, prepare lessons, simplify difficult concepts, translate terminology, create worksheets, assess written work and draft administrative documents. Some teachers also used AI-generated conversations, audio and images to make lessons more interesting and accessible. However, most had learnt to use these tools on their own. 

In many ways, teachers are moving faster than the education system that supports them. Not every teacher, however, is moving at the same pace. The study found that 80.5 percent of male teachers had used AI, compared with 71.3 percent of female teachers. Household and caring responsibilities limited the time female teachers could devote to professional learning. Older teachers were also more likely to be left behind. Many may need more personalised support delivered at a pace that helps them build confidence gradually.

These findings dispute a common assumption that providing the same training to everyone creates equal opportunity. It does not. A training opportunity is not genuinely accessible if its timing, location or format prevents some teachers from participating. The AI literacy gap is not simply a matter of talent, confidence or motivation. It is likewise shaped by time, opportunity and the availability of appropriate support.

If AI is to benefit every teacher, professional development must respond to these differences rather than overlook them. Access is only one part of the challenge. Teachers were also clear about the risks. They worried that AI could produce false or culturally insensitive information. They feared that students might use it to complete homework without thinking. Some were concerned that overdependence on AI could weaken creativity, curiosity plus independent judgement. They also raised concerns about deepfakes, cybercrime, privacy and reputational harm. 

However, awareness of less visible risks was more limited. These included algorithmic prejudice, discrimination, informed consent and the routine protection of students’ personal data. This is why formal guidance and professional development are now urgent. Teachers should not be expected to navigate such intricate moral questions alone.

The teacher-led classroom research offered another important insight. For eight weeks, 21 teachers experimented with AI in English, Mathematics, Science, Nepali and Social Studies classes from Grades two to ten. They found that AI could help them prepare lessons, create a wider range of activities, support multilingual learning, strengthen formative assessment and encourage reluctant learners to participate. But AI alone did not create these changes. Mentoring, reflection and adaptation mattered. Above all, teachers’ professional judgement mattered.

Perhaps the most valuable outcome was not simply that teachers learnt to use new tools. It was that they became researchers in their own classrooms. They tried new ideas, observed what happened, identified problems and adjusted their plans. This is an important distinction. Professional development should not simply train teachers to follow machine-generated instructions. It should empower them to question, test and adapt what AI produces.

The study points towards five immediate priorities for Nepal:

Priority 1: Interim guidance now, national standards next

Priority 2: An AI literacy pathway built into existing teacher development

Priority 3: Mentored classroom practice at scale

Priority 4: Connectivity-resilient infrastructure, targeted where it changes outcomes

Priority 5: Municipal capacity and the financing pathway

These priorities are closely connected. Guidance without training may remain unread. Training without classroom experimentation may stay theoretical. Experimentation without mentoring may leave teachers uncertain. Technology without appropriate infrastructure may deepen current inequalities. And none of these efforts will last without local capacity and sustainable financing.

AI must never be presented as a replacement for teachers. It cannot reproduce the empathy, trust, cultural understanding, creativity and moral responsibility that good teachers bring to their classrooms. A machine may generate a lesson plan, but it does not know the child who arrived at school hungry. It cannot recognise the hesitation of a learner who is afraid to speak. It cannot understand the social and cultural realities of every community. Nor can it take moral responsibility for the choices made in a classroom. Those responsibilities remain with teachers. But when teachers are prepared and properly supported, AI can become a valuable professional partner, one whose answers they examine rather than accept.

My grade four teacher had to leave his family for 10 months to continue his learning. Today, professional learning can reach a teacher through a device held in one hand. That is an extraordinary change. Yet access to a tool is not the same as access to important learning. A device cannot provide direction, professional confidence or ethical judgement by itself. 

Teachers must decide whether it is safe, relevant and appropriate for their learners. They need the confidence to question the directions when something seems wrong. They also need an education system that provides guidance when the route is unclear. Our education system must now support them.