Are teachers ready for AI in classrooms?, ETEducation

AI is already finding its way into classrooms, lesson plans and homework. But as schools move from experimenting with artificial intelligence to using it more regularly, a more difficult question is emerging: are teachers being prepared to make good decisions about AI, or simply being taught how to use the tools?

The distinction matters.

A recent AI Readiness survey by the Centre for Teacher Accreditation (CENTA), covering respondents across 16 countries in Asia and Africa, found that 98% believed AI readiness was important for teachers and schools. Yet only 23% could think of even one teacher or school they considered AI-ready.

The finding does not mean that only 23% of teachers are ready for AI. It points to something more fundamental: the education system itself is still struggling to define what an AI-ready teacher looks like.

That uncertainty is becoming harder to ignore as AI use spreads.

Teachers are increasingly using generative AI to prepare lesson plans, worksheets, assessments and teaching-learning material. Students are using it to ask questions, complete assignments, generate ideas and process information. Training platforms are adding AI-focused courses and tools, while schools are beginning to formulate their own approaches to responsible use.

But using an AI tool and being ready for AI are not the same thing.

The first problem is that use is easier to measure than judgement

The numbers coming from different surveys tell a similar story, even though they measure different things.

The AI readiness gap

*Figures are drawn from separate surveys by CENTA, Central Square Foundation and Bharti Airtel Foundation and are not directly comparable.

The Central Square Foundation’s BaSE 2025 survey, which covered 15,000 respondents across low-resourced settings, including 12,500 households and 2,500 teachers, also points to a gap between technology exposure and deeper capability.

Among teachers using EdTech, 51% reported receiving technology training or workshops. But only 9% said that this training focused on AI tools. At the same time, 79% wanted more technology training.

The message is not that teacher training is failing. It is that the content and purpose of that training may now need to change.

Bharti Airtel Foundation’s experience provides another indication of how quickly adoption is moving. Its survey found that 83.2% of teachers surveyed had used AI. Among those users, 34.5% said they used AI regularly, while 42.4% used it occasionally while verifying the outputs.

That distinction is important. Verification is already becoming part of teachers’ AI behaviour.

The Foundation has also been expanding AI-enabled resources through TheTeacherApp. In February 2026, it partnered with CK-12 Foundation to integrate more than 45 AI-enabled teaching tools into the platform.

Yet adoption alone cannot answer whether those tools are improving learning.

“AI may change how students find and process information, but education must continue to focus on helping them question, evaluate, apply and build upon that information. That is where the role of the teacher becomes even more important,” Nuriya Ansari, CEO, Bharti Airtel Foundation, told ETEducation.

That may be the most important distinction in the emerging AI debate: the question is shifting from whether teachers use AI to what they do with it.

Ramya Venkataraman, Founder and CEO, Centre for Teacher Accreditation (CENTA) Pvt. Ltd., made a similar distinction in a recent conversation with ETEducation. “The question is therefore not simply whether teachers are using AI. It is whether they understand how to use it meaningfully and responsibly in their teaching practice.”

CENTA’s work on an Education AI Readiness Index is attempting to formalise that distinction. Its framework looks beyond adoption to the competencies students need, the competencies teachers need to develop those capabilities in students, and how teachers themselves can use AI effectively. CENTA describes the forthcoming index as a way to assess whether education systems can use AI “meaningfully and responsibly”, rather than simply measuring adoption.

The real test begins when the teacher has to make a choice

The harder questions begin after a teacher knows how to generate a lesson plan or worksheet.

Should AI be used for this particular lesson?

Should the teacher accept the answer it produces?

What should students be asked to do themselves?

When does AI help a student think, and when does it allow the student to avoid thinking?

What information should never be entered into an AI system?

And what happens when an AI-generated answer sounds convincing but is wrong?

These are not technical questions alone. They are questions of judgement.

Central Square Foundation’s work on AI Samarth reflects this broader approach. Developed with the Wadhwani School of Data Science and AI at IIT Madras, the initiative focuses on foundational AI understanding, ethical awareness, critical thinking and practical application.

Gouri Gupta, Senior Director, EdTech & AI, Central Square Foundation, told ETEducation that the focus has to move beyond teaching students and teachers how to operate AI tools. It has to include understanding data privacy, bias, misinformation and the limitations of AI, while helping learners evaluate AI-generated information and decide when and how the technology adds value.

Its premise is increasingly relevant to teacher development: students and teachers cannot be passive recipients of AI. They need to understand both its possibilities and its limitations.

The shift is also visible in teacher-development programmes.

Sushma Raturi, CEO, Saamarthya Teachers Training Academy of Research (STTAR), told ETEducation that the organisation’s six-month AI certificate course, launched around two years ago, began with awareness and an understanding of the possibilities and implications of AI. But the larger challenge, she argued, is integrating AI into the way teachers think about learning.

One question captures that challenge particularly well:

“Is AI supporting the child’s thinking, or is it doing the thinking for the child?”

For schools, that question could become one of the most useful tests of AI readiness.

A teacher who can generate ten differentiated worksheets in seconds may be more efficient. But if students are no longer required to formulate questions, evaluate evidence or construct an argument, the technology may have improved the process while weakening the learning.

That is why AI readiness cannot be reduced to a list of tools.

AI readiness cannot be separated from pedagogy

This is where the conversation with Vetti Giri, Assistant Professor, Azim Premji University, becomes important.

Discussions around AI readiness can move quickly towards tool operation, prompt writing and evaluation of AI outputs. But those skills have limited value if teachers lack a strong understanding of how children learn and how knowledge is developed within a discipline.

The risk is that teachers become better technicians without necessarily becoming better teachers. Consider a science classroom teaching genetic inheritance.

One teacher may use sophisticated technology to demonstrate a concept quickly and efficiently. Another may use the topic to encourage students to observe patterns, frame questions, test ideas and construct explanations.

The first approach may accelerate the acquisition of information. The second may do more to develop scientific thinking. AI can support both. But it cannot decide which approach is educationally appropriate.

That decision belongs to the teacher.

The same applies across subjects. A language teacher needs to know whether AI-generated text is helping students understand writing or simply allowing them to bypass the writing process. A mathematics teacher needs to decide whether an AI-generated solution is being used to understand a problem or to avoid solving it. A social science teacher needs to help students interrogate sources rather than treating a fluent AI response as evidence.

In other words, AI literacy without pedagogical and disciplinary grounding can make teaching faster without necessarily making learning deeper.

This is why inquiry, problem solving, argumentation, pattern recognition and abstract thinking remain central to AI readiness. The irony is that the arrival of AI may make traditional teacher competencies more important, not less.

The change is bigger than the individual teacher

Another problem is that AI readiness is often treated as an individual teacher’s responsibility. But a teacher cannot be fully AI-ready if the school has no policy on acceptable use, no reliable infrastructure, no mechanism for sharing good practice and no leadership support.

Prof. Vineeta Sirohi, Professor & Head, Department of Educational Administration, NIEPA, brings this institutional dimension into the discussion.

NIEPA has been expanding professional development around artificial intelligence, including programmes examining AI integration in teaching, learning and research and the use of AI in creating educational content.

But the challenge becomes very different when viewed through the lens of educational administration. For a school leader, the question is not simply whether teachers have attended an AI workshop. It is whether the institution has created the conditions in which teachers can experiment, share practices, understand risks and make informed decisions.

Infrastructure also remains a basic constraint.

In schools where electricity, computers, connectivity or technical support are unreliable, discussions about sophisticated AI adoption can seem disconnected from classroom realities. That makes an equitable AI strategy different from simply distributing access to AI tools.

A hub-and-spoke approach, where central institutions build capacity and support networks that can reach schools with fewer resources, may therefore be more realistic than expecting every school to independently develop AI expertise.

The first step is not necessarily giving every teacher access to more tools. It is ensuring that teachers have the infrastructure, leadership and professional support to use the tools responsibly.

Teacher training itself has to move beyond the course

The traditional model of professional development is relatively straightforward: identify a skill, conduct a workshop, provide certification and move on. AI is unlikely to fit that model.

The tools change too quickly. More importantly, teachers encounter AI not as an isolated technology but inside everyday decisions about lesson planning, assessment, differentiation, feedback and student work.

That means AI professional development will have to become more continuous and practice-based. A teacher might learn how an AI tool works on Monday, try it in a classroom on Wednesday, discover that it produces unreliable content on Friday and then discuss the experience with colleagues the following week.

That cycle of experimentation, reflection and improvement is closer to what AI readiness requires.

CENTA’s work points in the same direction. Venkataraman has argued that technology can support core teacher competencies, but cannot substitute for them.

“Technology, including AI, can support teachers in all of these areas, but these remain very core teacher competencies,” she said in a conversation with ETEducation. That also changes what professional development should measure.

Course completion is easy to record. Competence is harder.

A school can know how many teachers attended an AI workshop. It may not know whether those teachers subsequently redesigned an assessment, identified a hallucinated AI response, changed a lesson because AI was inappropriate, or helped students use the technology more critically.

Those are much better indicators of readiness.

The hardest thing to measure may be readiness itself

The 98% versus 23% CENTA finding captures the measurement problem neatly. Almost everyone agrees that AI readiness matters. Far fewer can identify what readiness looks like in practice.

A useful framework therefore needs to go beyond “AI skills”.

An AI-ready teacher should be able to:

understand basic AI concepts and limitations;

use AI tools where they add genuine educational value;

verify AI-generated information rather than assuming it is correct;

recognise bias, misinformation and fabricated content;

protect student data and understand privacy risks;

design learning activities that preserve student thinking and agency;

use AI appropriately for planning, differentiation and feedback;

evaluate whether AI is improving learning rather than simply improving efficiency;

teach students how to question and evaluate AI outputs; and

know when not to use AI.

That last competency may ultimately be one of the most important. The best use of AI in a classroom may sometimes be no use of AI at all.

Equity cannot be an afterthought

There is another reason the teacher-readiness question matters. India’s schools operate in vastly different technological environments. A teacher in a well-resourced urban school may have access to high-speed internet, devices, paid AI platforms and technical support. A teacher in a remote government school may still be dealing with unreliable electricity, limited devices or poor connectivity.

The gap is therefore not simply between teachers who know AI and those who do not. It can also be between schools that have the conditions to experiment with AI and those that do not.

That is why AI-readiness programmes need to account for infrastructure, access and context rather than assume that every teacher starts from the same point. Otherwise, AI could deepen existing inequalities in teacher capability instead of narrowing them.

So what does an AI-ready teacher look like?

The answer emerging from conversations with teacher-development organisations, universities and education administrators is more demanding than a list of digital skills. An AI-ready teacher is not necessarily the teacher who uses the most AI tools.

It is the teacher who understands why a tool should be used, when it should be used, how its output should be evaluated and when it should not be used at all.

That requires AI literacy. But it also requires subject knowledge, pedagogy, critical thinking, ethical awareness and an understanding of learners. It also requires a school environment that gives teachers room to practise those skills. This is perhaps the biggest shift in the teacher-training conversation.

For years, technology integration in education was framed largely as an access problem: give teachers devices, connectivity and training, and adoption will follow.

AI is exposing the limits of that model.

The harder challenge is judgement.

The teacher’s role may become more human, not less

AI can increasingly produce explanations, questions, examples, worksheets, lesson plans and assessments in seconds.

That does not make the teacher less relevant.

It changes where the teacher’s value lies.

If information becomes easier to generate, teachers become more important as people who help students decide what information to trust, what questions to ask, what evidence matters and how knowledge should be used. That is also why the emerging AI-readiness debate should not be reduced to whether teachers are comfortable with technology.

The more important question is whether they are prepared to make educational choices in a world where technology can increasingly do things that teachers once had to do themselves. As Ansari put it, the task of education remains to help students “question, evaluate, apply and build upon” information.

The future-ready teacher, then, may not be the one who knows every new AI tool. It may be the one who knows when a tool improves learning, when it gets in the way, and how to make sure the student remains at the centre of the process.