The rise of generative AI has brought us to an inflection point in higher education. A recent Harvard Business Review analysis noted that over 60 percent of college students globally have already experimented with tools like ChatGPT for assignments or idea generation. Meanwhile, an MIT Sloan study highlighted that professionals using generative AI were able to complete knowledge tasks 40 percent faster and with greater accuracy compared to peers.

If such productivity gains are reshaping the workplace, it is natural that students, as future knowledge workers, will embrace these tools as well. Yet, in India, where the National Education Policy (NEP) 2020 is championing critical thinking and experiential learning, a pressing question remains unresolved:

Who owns the output of student work when it is generated or co-created with AI?

This is not merely a legal or philosophical curiosity. The answer will shape assessment fairness, academic integrity, intellectual property rights and ultimately, how India prepares its next generation for the AI-augmented workforce.

Why Ownership Matters in an Indian Context

India’s education system has long been designed around examination-driven evaluation. From CBSE boards to university exams, the implicit assumption has always been that submitted work reflects the student’s original effort. AI now destabilises this assumption.

Consider three real-world scenarios I’ve observed in my work with academic institutions and corporate learning teams:

1. The Shortcut Risk

A student feeds an essay prompt into a generative AI system and turns in the output verbatim. If the grade is awarded, does it reward the student’s skill, or the AI’s?

2. The Collaborative Augmentation

A student uses AI to brainstorm ideas, refine drafts or simulate different perspectives, but the final submission reflects their judgment and curation. In this case, shouldn’t ownership of the work remain with the student, much like citing research papers or software libraries?

3. The Commercial Extension

An engineering student builds an AI-assisted code module during coursework and later spins it into a startup product. Should the intellectual property belong fully to the student, or should the institution have a share if AI tools provided by the university were used?

Without a clear framework, universities and policymakers risk creating confusion and, worse, inequity. Wealthier students may access premium AI tools privately, while others depend on institution-provided access, blurring the fairness of evaluation.

Why We Need to Move From Policing AI to Cultivating Originality

The debate over copyright or disclosure, while important, is ultimately a distraction if our education system continues to reward reproducibility over originality. Generative AI thrives on well-posed, common problems because it has been trained on a vast diet of existing code, essays and ideas. But originality, the ability to frame a novel question, apply contextual judgment or integrate multiple perspectives, is where students must be pushed.

In India, this is where education policy needs to intervene. Instead of only asking, “Did the student use AI?”, the more transformative question is: “Did the assessment demand something only a thinking human could create?”

Consider two examples from software engineering education:

Problem Framing vs. Problem Solving

If an exam asks, “Write code to sort a list using merge sort,” ChatGPT or GitHub Copilot can deliver a textbook solution instantly. But if the assignment is reframed as, “Given data from a retailer’s POS system with inconsistent formats and missing fields, how would you design a data pipeline to ensure accuracy before applying any machine learning?”, suddenly, originality is unavoidable. The student must demonstrate domain understanding, judgment about trade-offs and creativity in framing the problem. AI can assist, but it cannot substitute solid reasoning.

Design Trade-offs in Real-World Systems

Instead of asking, “Implement a linked list in C,” consider asking, “If you were designing a payment gateway for small businesses in India, how would you balance low latency with fraud detection checks? Illustrate with pseudo-code and explain trade-offs.” Here, the value isn’t in raw code (AI can produce that), but in the why behind each decision. Students must connect technical architecture with business and societal context, something no generative model can hallucinate with reliability.

These are not futuristic hypotheticals. In India’s tech corridors, from Bengaluru’s startup hubs to large IT campuses in Hyderabad, I’ve seen young engineers who lean too heavily on AI copilots falter when asked to design solutions even slightly beyond the template. The true differentiators are those who can step back, frame the original question, and justify their design choices with clarity, skills that no machine can automate.

This is the direction India’s education policy must take: cultivate originality, not reproducibility.