Summary
Repeated question paper leak allegations in India have raised concerns about the security and reliability of public examinations. Infosys co-founder Nandan Nilekani has suggested that artificial intelligence (AI) could help address this problem by generating examination questions closer to the time of the exam.
The idea involves moving away from relying entirely on traditionally prepared question papers and exploring digital examination systems that can generate questions on demand. While this approach could reduce certain risks associated with paper leaks, implementing it across India’s large education system would require strong digital infrastructure, reliable security measures, and careful planning.
What Happened?
Question paper leaks have become a major concern for students, parents, educational institutions, and government authorities in India. Allegations of examination irregularities have prompted questions about how confidential papers are prepared, stored, transported, and distributed before examinations.
Against this backdrop, Nandan Nilekani has highlighted the potential role of technology in improving examination security. He suggested that AI could be used to generate question papers on the day of an examination, reducing the time during which a completed paper could be accessed or leaked.
The proposal points towards a possible shift from conventional paper-setting processes to more technology-driven examination systems.
However, AI-generated examinations would still require oversight to ensure that questions are accurate, fair, relevant to the syllabus, and consistent in difficulty.
How Could AI Help Prevent Question Paper Leaks?
Traditional examinations often involve preparing question papers in advance and keeping them confidential until the examination begins. If someone gains unauthorized access during this period, the contents may be leaked before students sit the exam.
AI could offer an alternative by generating questions from an approved question bank or syllabus shortly before an examination.
A possible AI-powered examination process could work as follows:
- Define the syllabus: Examination authorities establish the subjects, topics, learning objectives, and difficulty levels.
- Prepare a verified question bank: Subject experts create and review a secure collection of questions.
- Generate questions using AI: An AI system selects or creates questions according to predefined examination rules.
- Review and validate the paper: Automated checks and authorized human reviewers verify accuracy, relevance, and difficulty.
- Distribute the paper securely: Questions are delivered through a controlled digital system or printed under secure conditions immediately before the exam.
If implemented properly, this approach could reduce the time between question-paper creation and the examination, potentially limiting opportunities for advance leaks.
However, generating a paper on the exam day would not automatically eliminate leaks. The system would still need protection against unauthorized access, insider threats, cyberattacks, and the theft of generated questions.
Why Digital Examinations Could Make a Difference
Digital examination systems could provide additional security features compared with conventional paper-based processes.
For example, authorities could restrict access to question papers, maintain audit logs, use encryption, and generate different question sets for different examination centres. Questions could also be released through controlled systems at a scheduled time.
AI could help select questions with comparable difficulty levels, allowing different students or examination centres to receive different versions of the same test.
Such systems may also reduce dependence on transporting large quantities of printed question papers across different locations.
Nevertheless, digital examinations would introduce their own challenges, including internet connectivity, power failures, device availability, cybersecurity, and the need to prevent cheating during online tests.
Challenges of Implementing AI in India’s Education System
Although the proposal offers potential benefits, implementing AI-powered examinations nationwide would be a substantial undertaking.
1. India’s Large Examination System
India conducts numerous school, university, entrance, recruitment, and professional examinations involving millions of candidates. Building a secure digital system capable of handling this scale would require considerable investment and coordination.
Not every examination centre has the same level of digital infrastructure, making a uniform rollout difficult.
2. Accuracy and Fairness of AI-Generated Questions
AI systems can sometimes produce incorrect, ambiguous, repetitive, or misleading questions. In high-stakes examinations, even a small number of errors could affect students’ results and future opportunities.
Subject experts would need to validate questions, while authorities would have to ensure consistent difficulty levels across different versions of an examination.
3. Cybersecurity Risks
Moving examinations to digital platforms would not remove security risks; it would change their nature.
Hackers or unauthorized insiders might attempt to access question banks, compromise examination software, or exploit weaknesses in the systems used to generate and distribute papers.
Strong encryption, access controls, independent security audits, and clear incident-response procedures would be essential.
4. Infrastructure and Accessibility
Digital examinations require suitable devices, reliable electricity, secure networks, and trained staff. Students in areas with limited infrastructure could face disadvantages if the transition is not planned carefully.
Authorities would need accessible examination facilities and reliable backup arrangements for technical failures.
Can AI Replace Human Question Paper Setters?
AI could assist with generating questions, organizing question banks, checking difficulty levels, and identifying possible errors. However, it should not necessarily replace subject experts and examination authorities.
Human oversight remains important for evaluating whether questions test the intended knowledge and skills, whether they contain cultural or language bias, and whether they follow the official syllabus.
A practical approach would combine AI-assisted question generation with independent expert review and secure examination procedures.
This would allow authorities to benefit from automation while retaining accountability for the quality and fairness of public examinations.
Nandan Nilekani’s View on AI and Jobs
Nilekani has also expressed the view that artificial intelligence will change the employment landscape, while creating opportunities for people who continue to update their skills.
As AI tools become more capable, some routine tasks may be automated, requiring workers to adapt to changing job responsibilities. At the same time, AI adoption may create demand for professionals with skills in software development, data analysis, cybersecurity, AI implementation, and other technology-related areas.
The broader message is that continuous learning and the ability to work with emerging technologies may become increasingly important across industries.
However, the impact of AI on employment will vary by sector, occupation, and the pace at which organizations adopt the technology.
What Happens Next?
AI-generated question papers remain a proposal that would need detailed evaluation before being adopted for major public examinations.
Education authorities would need to assess whether AI can produce reliable questions at scale, whether different versions maintain comparable difficulty, and whether digital delivery offers meaningful security advantages over existing processes.
Pilot programmes could help test the approach before wider implementation. These trials should measure question quality, system reliability, accessibility, security, and the cost of operating the platform.
AI could become one part of a broader examination-security strategy, alongside better confidentiality procedures, stronger monitoring, independent audits, and stricter action against proven misconduct.
Conclusion
Artificial intelligence could offer India a new way to improve examination security by helping generate question papers closer to the examination date. As suggested by Nandan Nilekani, digital systems and AI-assisted question generation may reduce some of the risks associated with preparing and distributing confidential papers in advance.
However, technology alone cannot guarantee leak-free examinations. Success would depend on secure infrastructure, accurate questions, expert oversight, fair access, and effective monitoring.
Rather than treating AI as a complete solution, India could explore it as part of a wider effort to make public examinations more transparent, reliable, and secure.