OPINION | AI in Classrooms: Necessary reform or premature distraction?

Challenges of AI in the Classroom

There is a widespread belief that AI will disrupt the lives of every citizen – and perhaps already is. This naturally raises the question: if AI is changing the world, how do we prepare our children for it?

India has responded by introducing “AI and computational thinking” from Grade III onwards. The idea is seductive – AI as a universal future skill, but this top-down push sits uneasily with the realities on the ground. We are still grappling with gaps in basic literacy and numeracy – an admission by NEP 2020, with the country embarking on a universal functional literacy and numeracy mission. Will early AI classes help close these gaps, or distract from them?

What does introducing AI to children even mean? Broadly speaking, there are three possibly overlapping approaches:

1) Use of AI tools in schools – Curriculum and pedagogy are structured to include the use of AI tools in classrooms, with teachers guiding their use in particular ways.

2) Studying AI’s impact on society – Children study the effects of AI on society and develop a critical perspective on its role.

3) Teaching the internals of AI – Children are taught to “look under the hood,” learn the internal structure of AI, and develop mastery over the technology.

The Challenges of AI in the Classroom

Introducing newer pedagogical tools in the classroom is always welcome, but the nature of AI is fundamentally different from that of a calculator or a smartboard. It interferes with cognitive processes – by supplying answers, shaping writing, or proposing arguments – which can render children vulnerable. This places a considerable burden on the teacher and the school, and it is essential to train them adequately on the benefits and, more importantly, the risks of using AI before deployment.

Subjecting the relationship between digital technologies and society to critical scrutiny is certainly an important educational agenda. However, this requires considerable understanding of how technology functions and considerable maturity, which is why it may not be an appropriate theme for children.

Is Teaching AI the Right Approach?

That leaves the third approach – the one that the Government favours – teaching AI in schools. This is presumably to universalise AI skills and propel India towards technological and economic leadership. Unfortunately, technology education is complex and is universally taken up only at the tertiary level, with basic introductions at the secondary level. Even traditional apprenticeship models typically start at the upper primary stage.

One argument is that children will eventually need to learn AI – the earlier, the better. But consider two simple facts:

# AI is not (yet) a compulsory core subject across disciplines even in universities, so it is not obvious that it should be a part of the core compulsory curriculum in schools.

# Technological disruptions are not new. The internet and mobile phones have impacted every citizen’s life. Citizens of all ages have had to master the use of these technologies to succeed in their fields, even to function in life and society. They did not have to learn the internal structure of the internet or the underlying technology of mobile phones. School or university curricula were not modified to teach these technologies.

The Age-Appropriateness of AI Education

What is to be taught to children at different ages is decided by considerations of what educational purpose is served, and what is appropriate at each stage of cognitive, social, and emotional development. Young brains develop learning pathways best through concrete experience and play. Algebra is immensely useful, but you would not teach algebra to a kindergartner who can’t add yet, would you? The complexity of concepts underlying AI makes it inappropriate for children’s learning. On the other hand, “dumbing it down” simply to teach AI to children serves no purpose.

Potential Harm: Overloading Cognitive Skills

On the other hand, imposing programming concepts or machine learning ideas on primary school children can do great harm. It may overload working memory and confuse more than clarify. Research in cognitive science warns of just this danger. Studies of AI tools in education note that handing students answers via AI may “limit critical thinking skills,” creating knowledge pipelines that students accept uncritically. One analysis warns that students who lean on AI too much can end up with “weaker problem-solving skills.” In plain terms: if a child learns to always ask a machine for the answer, they may never learn to ask good questions themselves.

Indeed, the advent of AI should strengthen the agenda for critical thinking in our school classrooms, rather than focusing on teaching AI in schools. As the impact of AI deepens, there will be bot-assistants for artists, scientists, and other professionals. This only means that our children will need to develop a deeper understanding of the arts and sciences (than we do currently) so that they can effectively use AI, and not surrender to its capabilities.

The Need for Strong Conceptual Foundations

In the near future, human beings will need to be better in their own areas of competence than AI bots. Indeed, they will need to learn to master AI agents that are not yet invented. This will require them to have strong conceptual and cognitive foundations in their domains of work, and the ability to adapt to new technologies.

School has the all-important task of nurturing and developing children’s innate intelligence. This is as yet a work in progress in our schools. Let the seductive technology of artificial intelligence not interfere with this.

(R Ramanujam is a Faculty Member at Azim Premji University and R Venkatesh, formerly with TCS research.)

Views are personal, and do not represent the stance of this publication.