Generate questions at last minute to prevent paper leaks: Nilekani

Generating a paper at the last minute means officials would not prepare one complete, final paper weeks or months before an examination. Instead, they would select questions from a large, secure bank shortly before candidates begin. The aim is to remove the long period during which a finished paper could be copied, stolen, or sold.

For example, the system could choose a required number of questions from different subjects and difficulty levels. It could then create a paper for that specific examination session. Nilekani also said artificial intelligence could generate papers on examination day. The central mechanism is timing: the final combination would exist only briefly before the test.

The proposal is part of a wider examination-security overhaul. The task force is considering technology, better testing centres, and separate responsibilities among agencies. It is preparing its report after receiving more than 50,000 suggestions. The plan would require secure systems and careful controls over the question bank.

A question bank would contain many approved questions covering the subjects and skills an examination must test. Just before the exam, software could select a specified number from each required category and assemble them into a complete paper. This would allow different sessions or candidates to receive different combinations.

For instance, a selection system might draw questions from several topics while following preset rules for coverage and difficulty. It could then deliver or print the final paper immediately before the examination starts. The article suggests that artificial intelligence might assist with this process on examination day. The important safeguard would be a secure, controlled selection process.

This approach is proposed because a fixed paper creates an economic incentive for theft and sale. It is not yet described as an implemented system. The six-member task force is preparing recommendations on examination security, technology, governance, and the separation of examination responsibilities.

A fixed question paper has value before the examination because it contains the exact questions candidates will face. Anyone who obtains it early could sell access to candidates, coaching groups, or intermediaries. That creates a financial reason for someone with access to copy, steal, or leak the document.

Nilekani described this directly: once a question paper exists, it creates an economic incentive for it to be sold. Under his proposal, questions would come from a massive bank and the final paper would be assembled just before the exam. The shorter the exposure period, the less useful advance theft becomes.

This idea forms part of a broader technology-driven reform effort. The task force is also considering secure computer-based testing centres and separate agencies for paper-setting, exam conduct, centre management, and security. Its report is expected later this month, after consultations and more than 50,000 suggestions.

The article says India’s existing testing network includes around 1.5 lakh testing locations. That is the scale of infrastructure involved in moving more high-stakes examinations from paper to computers. The network would need to support candidates, secure systems, reliable connectivity, and examination-day operations.

Nilekani stressed that this network must expand to accommodate a wider shift toward computer-based examinations. The task force’s suggestions also include better-equipped centres with reliable electricity, internet connectivity, biometric verification, and CCTV surveillance. These facilities would help examinations run consistently and securely across many locations.

The figure describes the current network, not a confirmed target for expansion. The task force is still preparing its report after around 18 meetings and 40 consultations. Its work follows the NEET-UG controversy and student protests, and includes proposals on technology, governance, multiple attempts, and examination security.

Debundling means separating the main examination responsibilities. One agency could set questions, another could conduct the exam, another could manage testing centres, and another could handle security. The proposal seeks to reduce concentration of control over the examination process.

For example, the paper-setting agency would not also control centre operations or security arrangements. Independent teams could perform different checks and limit how much information any one group holds. If one part of the process faced a problem, responsibility would be easier to identify and the entire system would not depend on one organisation. The article presents this as a way to reduce risk.

The proposal is under consideration by the task force, not described as completed. The panel has held around 18 meetings and 40 consultations. Its wider agenda includes last-minute paper generation, computer-based testing, secure centres, and a possible common calendar for major recruitment and entrance examinations.

Last-minute paper generation could reduce the opportunity to obtain and sell an exact paper before the examination. Biometric checks could help centres verify that the person taking the test is the registered candidate. CCTV surveillance could provide monitoring of activity at examination locations. Together, these measures address different parts of examination security.

For example, a candidate might be required to complete biometric verification before entering a computer-based test centre. The paper itself could then be assembled securely just before the session begins, while cameras monitor the centre. The mechanism is layered protection: fewer people may know the final paper, and centre access and activity can be checked.

The article presents these as recommendations or suggestions received by the task force, not as confirmed nationwide practice. Other proposals include reliable electricity, internet connectivity, multiple attempts, and better-equipped centres. The panel is expected to submit its report later this month after extensive consultations.

Computer-based exams can give different candidates different questions while keeping tests comparable. The key is to build a question bank with questions that have been reviewed and calibrated for content, skill, and difficulty. A selection system then follows the same blueprint for every candidate or session.

For example, every test might include the same number of questions from each subject area and several difficulty levels. Questions can be compared using performance data from earlier trials. Statistical equating can then adjust scores when separate forms are not exactly identical. This makes the result depend on demonstrated ability, rather than on receiving an unusually easy or hard set.

The article focuses on generating papers from a massive bank but does not describe fairness controls or scoring methods. Those safeguards would therefore need to be specified in the task force’s recommendations. Nilekani’s broader proposal also requires expanded computer-based testing infrastructure and secure centres.