Dr Akhil Damodaran, Director, Symbiosis Centre for Management and Human Resource Development (SCMHRD), Pune.
A student submits an impressive business strategy assignment. The market analysis is comprehensive, the financial projections are convincing, and the recommendations are logically structured. Generative AI may have helped the student research, analyse and present the work.
But what if the next stage of assessment involved a simple question: "Suppose your largest customer withdraws tomorrow. Would you still recommend the same strategy?"
The student's response could reveal something equally valuable as the assignment itself: the ability to understand, defend and apply knowledge.
As artificial intelligence becomes increasingly integrated into education, perhaps we should explore using it not only to help students prepare assignments but also to assess their understanding.
My proposition is simple: let students use AI responsibly to prepare their work, and then let an AI examiner question them individually about what they have submitted.
At Symbiosis Centre for Management and Human Resource Development (SCMHRD), Pune, we have been experimenting with AI across management education, including business simulations, learning platforms, strategic decision-making exercises and academic evaluations.
One particularly interesting experiment involved an AI-powered Summer Internship Project evaluator.
A student had completed an internship with a pharmaceutical company. The project involved specialised industry processes and technical concepts. Our management faculty examined the methodology and managerial implications, while certain technical aspects offered an opportunity for additional sector-specific exploration.
We invited the student to present the project before SCMHRD's AI evaluator.
The system generated pharmaceutical-sector questions, explored technical explanations and encouraged the student to clarify underlying assumptions. The interaction demonstrated how AI could introduce additional disciplinary perspectives into academic evaluation.
This was an exploratory exercise, not proof that AI could independently verify every technical claim. Its questions and judgements still required faculty oversight. Nevertheless, it suggested an interesting possibility: AI could expand the questioning capacity of academic assessments while complementing faculty expertise.
Our experiments with AI-supported boardroom and business simulations point in a similar direction.
Students make strategic decisions while responding to changing market conditions, competitive pressures and organisational priorities. A recommendation that appears appropriate initially may require reconsideration when circumstances change.
Why not introduce similar interactions into examinations?
Imagine a student submitting an AI-assisted market-entry strategy. An AI examiner could ask the student to justify the market selection, explain financial assumptions and reconsider the strategy when customer acquisition costs suddenly increase.
The objective would not be to identify mistakes, but to encourage students to demonstrate reasoning, adaptability and practical understanding.
This distinction is increasingly relevant in the professional world.
AI can generate financial models, business proposals and strategic recommendations. However, organisations also need professionals who can interpret these outputs, understand their limitations and exercise informed judgement.
Higher education should prepare students for both capabilities.
AI-assisted assessment could make individual academic interactions more accessible. Students could participate in structured preliminary vivas, while faculty review responses, provide feedback and retain responsibility for academic evaluation.
A practical starting point could be to allocate 20-30% of the overall assessment weightage to AI-assisted evaluation, with the remaining weightage continuing through established faculty-led methods. This would allow institutions to experiment, compare outcomes and refine the approach based on academic evidence and feedback.
Such an approach need not be introduced across every course immediately. Institutions could begin with selected assignments, projects and case-based assessments, gradually expanding its application as confidence in the process develops.
The approach would also require thoughtful safeguards. Faculty must define learning outcomes, questioning frameworks and evaluation criteria. AI-generated assessments should be reviewed for accuracy, fairness and consistency, with appropriate protections for student information and confidential industry data.
Equally important, students should know the rules in advance, including which AI tools they may use and which competencies they are expected to demonstrate.
Our experiments at SCMHRD represent early explorations of these possibilities. They suggest that AI can be considered not only as a learning companion or productivity tool but also as a questioning partner that enriches the assessment process.
For Indian higher education, the opportunity is significant. Rather than viewing AI-assisted assignments and meaningful assessment as competing priorities, institutions can explore how the two complement each other.
Students can use intelligent tools to research, develop ideas and prepare recommendations. AI-assisted examinations can then invite them to explain their decisions, explore alternative perspectives and demonstrate independent understanding.
Faculty remain central to guiding learning and maintaining academic standards.
Ultimately, education should develop graduates who can work confidently with AI while retaining the ability to think critically, challenge assumptions and make responsible decisions.
If AI can write the assignment, let AI also be the examiner. Let students use AI to prepare their work, and then challenge them to demonstrate their understanding when AI asks the questions.
Perhaps the next step in education is not to make students compete against AI, but to let AI challenge their thinking.
