Shaping Tomorrow’s Classrooms: Education’s Tryst with AI
When computers were introduced in my school, they were placed in special rooms—air-conditioned “Computer Labs”. We had to leave our footwear outside, before we entered, reminding us that this was a special and a sacred space, where miracles happened. The machines had bulky CRT screens, with blurry monochrome lines, and operated with floppy disks. When we switched from the 5 inch floppy disk versions to the sturdy 3 inch ones, it seemed like a generational leap. Wide eyed, we learnt about applying our first algorithms—Fibonacci series and prime numbers, and translated them to computer programmes. That was my window to the world of technology, where now Artificial Intelligence (AI) dominates and my school computer science education seems like centuries away.
AI is now everywhere. Use of AI applications and services are all around us and are increasing rapidly. In business, a recent McKinsey survey of organisations showed that 78 percent of organisations use AI in at least one business function, up from 72 percent in early 2024 and 55 percent in 2023. This trend is evident in the use of AI in higher education as well. 86 percent of students use AI for studies as per the survey by the Digital Education Council across 16 countries in 2024.
We use AI everyday, either with full awareness of the use case, for e.g., through large language models such as ChatGPT, or without awareness of AI, when it operates in the background to manage recommendations for us, such as YouTube, or on social media. AI is deepening its roots across all sectors, especially in knowledge-based work—we conduct research, generate emails, reports, images, and other creative artefacts using AI, in a matter of minutes.
My generation of millennials are busy adapting to this new AI world, as we must. The more important questions to ask are: What does this mean for the generations to come? How should they learn about and benefit from the proliferation of AI? How do we prepare them for a future where AI is integral to all elements of work and life?
AI in education has progressed rapidly, especially in schools (K-12). For simplicity, we can divide this into two specific segments:
- What we learn: Learning about AI and how to use AI applications for students and teachers
- How we learn: Use of AI in improving efficiency, and effectiveness of learning, i.e., personalised learning and better outcomes
1. What we learn
Across the world, AI is getting integrated into the curriculum for students and teachers. It is critical that we drive responsible integration of AI, blending curriculum innovation, teacher empowerment, and strict data privacy compliance. The goal is to prepare children to use AI and to critically engage with it.
Select examples of country efforts are outlined below (source: DevelopmentAid):
- Australia has implemented the National Framework for Generative AI in Schools beginning in 2024
- Estonia’s KrattAI initiative aims to ensure all students aged seven to 19 achieve digital fluency by 2030, with particular focus on ethical AI application and identifying algorithmic bias
- The UAE government has just approved the introduction of Artificial Intelligence as a subject across all government schools, starting next academic year 2025-26
- China has mandated AI integration into national curricula, textbooks, and teaching practices from September 2025
- South Korea has rolled out AI-powered digital textbooks for mathematics, English, and computing in March 2025
- In India, the National Education Policy (NEP) 2020 emphasises the integration of emerging technologies, including AI. State boards and private schools are adopting AI curricula, and there is a growing ecosystem of edtech for K-12 AI education. Initiatives such as AI Samarth by Central Square Foundation, supported by Google.org, are set to scale knowledge of AI, encouraging the responsible use of AI among students, parents, and teachers across India with the goal of reaching 5 million students. Moreover, the push for AI integration is supported by digital platforms like DIKSHA and SWAYAM supporting teacher training and content dissemination
5. How we learn
AI applications in K-12 education promises to reshape how educational content is created, delivered, and personalised.
Content development: User generated or public interest content combined with edtech was transformative for education in the last decade. The next wave of edtech is generative AI. Developing text and multimedia content and teaching aids will offer a superior pedagogy. Moreover, there is an opportunity to meet the requirements of diverse learning needs in classrooms, especially for children with learning differences. For e.g., The U.S. Department of Education has identified several areas where AI enables educational priorities with efficient content development, and support for teachers with equity and accessibility enabled by responsible use.
Personalised Teaching and Learning: Intelligent Tutoring Systems (ITSs) or Personalised Adaptive Learning (PAL) is a well researched AI application in K-12 education. One well-known example of an ITS is Duolingo, a mobile application for language learning that personalises instruction for each user.
A paper on “A systematic review of AI-driven intelligent tutoring systems (ITS) in K-12 education”, which analysed 28 studies, found that ITSs generally produce positive effects on learning and performance. DIKSHA (Digital Infrastructure for Knowledge Sharing), an Indian government initiative seeded by EkStep, has integrated PAL. It uses AI to customise learning opportunities for individual students based on their strengths, weaknesses, and learning styles. It offers adaptive content, real-time feedback, and personalised curriculum paths. AI helps monitor student progress and provides data-driven insights for teachers.
Evaluations, grading, and remedials: AI can be used for revamping assignment evaluations with personalised and unbiased feedback, which can feed into adaptive learning. Moreover, remedials and doubt clearing with always-on personalised assistance can be effective for students and teachers in improving learning outcomes.
Moving forward: The field of education is no stranger to the use of technology. AI will scale rapidly for both students and teachers on both dimensions of what they learn and how they learn. However, we need a laser focus on the following interventions, so that existing inequities are not exacerbated.
- First, the gender digital divide has widened with increasing internet and mobile penetration. More people access the internet than ever before, however, the proportion of women with digital access is not improving. Women in low- and middle-income countries (LMICs) are 14 percent less likely than men to use mobile internet, equating to 235 million fewer women using mobile internet than men. As we design for AI in education, we need to be intentional about reaching more girls in school with education of AI and empower them with devices and tools that can be their gateway to equity.
- Second, data protection and privacy for children needs to be built in the design and cannot be a check in the box exercise. For e.g., in India, despite limited exemptions for educational purposes, processing children’s data (age under 18) requires verifiable parental consent, strict age and identity checks, and prohibits tracking, behavioral monitoring, and targeted advertising. Similarly, the Children’s Online Privacy Protection Act (COPPA) in the U.S., GDPR across the EU and several other countries place restrictions on collection, processing and use of Children’s data. As we think about personalised learning through AI, building robust guardrails is critical.
- Third, we need tools and standards to assess safety of AI in edtech through standardised responsible AI frameworks. As edtech solutions emerge across curriculum and adaptive learning, it will be useful for parents and teachers to assess the safety and efficacy of these tools. The SAFE Benchmarks Framework (EDSAFE AI Alliance) – a coalition steered by a Brooklyn nonprofit, InnovateEDU, is driving policy guidance and benchmarks from the US. The UK government has published safety expectations for generative AI products used in education, requiring robust content filtering, age-appropriate moderation, technical mitigations for identified risks, and ongoing testing with diverse user groups. International bodies like the OECD and G7 have developed actionable, school-specific governance frameworks to guide risk assessment, transparency, consent, and alignment of AI tools in K-12 settings. As an example from India, the EdTech TULNA, a public good, established the standards for the design of edtech products across content quality, pedagogical alignment, and technology and design. Similarly, we need public goods that focus on AI education, similar to standardised RAI evaluations with benchmarks like HELM Safety, AIR-Bench, and FACTS that offer tools for assessing factuality and safety of AI models.
AI in education is chugging ahead at pace. We have the opportunity to build the future for generations to come with intentionality. We are not using floppy disks anymore!