Research agenda

Nothing here is a promise that any of it will be done. It is published anyway, for three reasons: it shows what an exam-only answer leaves untouched; it lets anyone see which question their own experience speaks to; and it is an open invitation — pick a section, and it becomes yours to work on.

Each workstream must begin with Indian variation — government and private, rural and urban, wealthy and poor, across states, languages, castes, genders, disabilities, and ages — before borrowing a fashionable model from elsewhere.

1. Purpose and the learner

• What should everyone learn for work, daily life, relationships, citizenship, creativity, and further learning?
• How do people discover a meaningful pursuit when interest, aptitude, exposure, family expectations, labour markets, and growing competence interact?
• How can a system cultivate excellence without treating most people as the failed remainder of a rank order?

2. Curriculum and syllabus

• What is foundational for everyone, what should be elective, and when?
• What should be removed to create time for depth, projects, arts, movement, making, discussion, and reflection?
• Which knowledge remains essential when retrieval and fluent generation are cheap, and which new literacies — statistical, media, computational, civic, financial, ecological, and AI — become foundational?
• How can national coherence coexist with state, language, local, indigenous, and learner choice?

3. How learning happens

Compare lecture and direct instruction, dialogue, deliberate practice, mastery progression, projects, apprenticeship, play, peer learning, mixed-age learning, outdoor work, studios, laboratories, and self-directed study. Ask what works, for whom, for which domain, at what age, and under which classroom constraints.

Khan Academy's mastery model and distinctive schools in Silicon Valley and elsewhere are cases to investigate, not brands to celebrate. Include ordinary public systems that achieve strong and equitable outcomes, Indian innovations, and experiments that failed or selected unusually advantaged families.

4. Teachers and classroom reality

• Measure actual teacher time: teaching, preparation, feedback, administration, reporting, election duties, compliance, pastoral care, and training.
• Reduce non-teaching load before prescribing another programme.
• Study class size, multilingual classrooms, subject knowledge, autonomy, career progression, peer learning, coaching, and psychological safety.
• Evaluate technology by whether it returns time and judgment to teachers.

5. Assessment and examinations

• Separate formative feedback, school completion, certification, selection, diagnosis, and system accountability; no single test can do all six well.
• Examine question design, validity, reliability, security, logistics, accessibility, grievance redress, retakes, modular assessment, portfolios, practical work, oral defence, open-resource tasks, and authentic performance.
• Model how every metric changes teaching, coaching markets, family behaviour, fraud incentives, and learner wellbeing.

6. Opportunity, affirmative action, and mobility

• Begin from constitutional equality, anti-discrimination, substantive justice, and the documented structure of disadvantage — not from slogans for or against reservation.
• Examine caste, tribe, class, income, wealth, gender, disability, language, region, school quality, first-generation status, and intersecting barriers.
• Compare instruments: early investment, weighted funding, outreach, preparation, reservations, contextual admissions, bridge support, scholarships, anti-discrimination enforcement, and transparent outcome audits.
• Test legal validity, political incentives, gaming, stigma, creamy-layer and exclusion questions, administrative feasibility, and effects after admission.

No reform should trade away the rights of disadvantaged groups under the guise of technical neutrality. No existing instrument should be immune from evidence, constitutional review, or better complementary design.

7. Pathways through life

Create permeable routes among school, vocational learning, apprenticeships, higher education, research, art, sport, entrepreneurship, care work, employment, and adult reskilling. Study credit transfer, recognition of prior learning, career guidance, financing, geographic mobility, and social prestige.

8. The pre-AI settlement

First repair known failures that do not require AI: early nutrition and health, foundational learning, libraries, teacher capacity, safe schools, disability access, language support, relevant curricula, laboratories, counselling, feedback, and institutional accountability. AI must not become an excuse to skip electricity, people, books, or rights.

9. The post-AI settlement

• Decide when AI is tutor, tool, collaborator, simulation, accessibility aid, translator, teacher assistant, assessor, or prohibited shortcut.
• Redesign assignments so learners must frame problems, show process, verify outputs, connect ideas, make things, defend choices, and exercise judgment.
• Protect privacy, childhood development, autonomy, authorship, and the right to learn without surveillance or a commercial profile.
• Prepare teachers and learners to understand model limits, bias, provenance, uncertainty, security, and environmental and labour costs.
• Preserve human relationships, productive struggle, embodied experience, community, art, play, and care as educational goods rather than inefficiencies.
• Prevent unequal access to high-quality human teaching and AI from widening existing advantage.

10. Institutions, law, finance, and politics

Map who has authority, who pays, who supplies, who audits, who profits, and who can appeal. Study centre–state responsibilities, boards and regulators, school and university autonomy, procurement, data governance, private coaching and edtech markets, corruption, regulatory capture, judicial constraints, and implementation capacity.

Every proposal needs a transition path that survives elections, bureaucracy, budget limits, federal variation, adversarial actors, and ordinary human error.

11. Comparative atlas

For each case — from Khan Academy and experimental schools to national systems, community schools, apprenticeships, and Indian state initiatives — record:

1. the problem and population;
2. the pedagogy and institutional model;
3. independent outcome evidence and time horizon;
4. admissions and selection effects;
5. teacher conditions and total cost;
6. equity, privacy, and wellbeing effects;
7. failures and criticism;
8. cultural, legal, fiscal, and scale constraints; and
9. what India could test reversibly rather than import wholesale.

12. Measures of success

Track foundations and advanced capability, curiosity and agency, wellbeing and belonging, creative and civic work, transitions into meaningful livelihoods, mobility across social groups, teacher time and retention, accessibility, recovery after failure, and long-run contributions to communities.

No dashboard captures a whole life. Use measures as evidence, not as the goal.