CBSE’s OSM Crisis: Building a trusted digital evaluation system

Drawing on the transparency framework adopted in MCQ-based examinations, there is an urgent need for National Digital Evaluation Model (NDEM) to resolve the OSM crisis and build a trusted digital evaluation system for Gen-N

Teaching-Learning-Evaluation (TLE) is the core of any education system, with evaluation as the most consequential component because it directly affects students’ academic progress and future opportunities. After years of effort, preparation, and investment, students have a legitimate right to know how their answer scripts were assessed-how marks were awarded, how step-marking was applied, which responses were left unchecked, and where marks were deducted despite substantial conceptual correctness. Transparency in evaluation is therefore essential for academic fairness and trust. Over the years, concerns have repeatedly surfaced about hurried evaluation, unchecked responses, inconsistent marking, examiner subjectivity, and unexplained variations in awarding marks, highlighting the need for more transparent, verifiable evaluation systems.

Responding to long-standing demands for greater transparency, CBSE introduced the On-Screen Marking (OSM) system for the Class XII Board Examination, replacing conventional pen-and-paper (PnP) evaluation with digital assessment. Physical answer books are scanned and evaluated digitally through a controlled software platform. While this marked a major shift towards transparency and standardisation, the rollout was soon overshadowed by technological glitches, evaluation anomalies, and concerns about the credibility and verifiability of the process.

Broadly, written examinations are of two types: Multiple Choice Question (MCQ)-based (objective) and descriptive (subjective). In both cases, evaluation systems are increasingly expected to leverage digital technologies to improve transparency, standardisation, accountability, and verifiability.

In MCQ-based examinations, evaluation is objective and straightforward because responses are matched against predefined answer keys. Such examinations commonly use Optical Mark Recognition (OMR) sheets, copies of which are often made available to candidates. After provisional answer keys are released, stakeholders may submit objections before final answer keys are frozen. Candidates can then independently estimate and verify their scores against declared results. Since the process is binary and machine-verifiable, it involves minimal subjectivity and limited scope for disputes.

This transparency framework-comprising disclosure of OMR sheets, publication of provisional answer keys, stakeholder feedback, and release of final answer keys-was pioneered and advocated by the author through sustained efforts during the early 2010s. It has since been adopted by most major admission and recruitment examinations in the country, such as IIT, CBSE, NTA, and others. More recently, UPSC adopted the same SOPs for the Civil Services (Preliminary) Examination, 2026.

Beyond transparency, the digital MCQ-OMR ecosystem has also provided valuable audit trails for investigative and judicial scrutiny, including in major admission and recruitment controversies such as the West Bengal teacher recruitment case.

Descriptive examinations operate in a fundamentally different environment. Answers are analytical, interpretative, or procedural, with marks awarded for intermediate steps, reasoning, methodology, diagrams, and presentation rather than a simple right-or-wrong outcome. Consequently, even under a common marking scheme, different evaluators may legitimately award somewhat different marks to the same answer.

This inherent subjectivity necessitates mechanisms for re-checking and re-evaluation. Re-checking verifies assessment and totalling accuracy, while re-evaluation involves a fresh assessment when significant discrepancies are suspected. However, in examinations involving millions of answer scripts, these processes are time-consuming and resource-intensive, and are typically invoked only when serious anomalies are suspected.

Encouraged by the success of digital evaluation in MCQ-based examinations, CBSE reportedly experimented with On-Screen Marking (OSM) on a limited scale from 2013-14 onwards. The nationwide rollout in 2026, however, represented one of India’s largest digital evaluation exercises, involving about 17 lakh Class XII students and nearly 98 lakh answer scripts. Such an undertaking required seamless integration of scanning infrastructure, secure storage, high-capacity servers, evaluation software, trained examiners, moderation systems, and student-access platforms. Given the scale and complexity of the exercise, the rollout appears to have proceeded without adequate large-scale dry runs, stress testing, and stakeholder preparedness, exposing significant implementation challenges.

However, the rollout was soon marred by reports of portal glitches, payment gateway failures, server congestion, and difficulties in accessing scanned answer scripts. Concerns deepened further over blurred scans, missing pages, unchecked responses, inconsistent marking, and questions regarding the completeness and correct indexing of digitised answer books.

The controversy showed that digital evaluation is far more than digitising answer books. It requires secure infrastructure, trained examiners, quality controls, and transparent verification. Failure at any stage can undermine trust in the system. Ironically, a reform aimed at enhancing transparency and accountability itself raised questions about transparency, verification, and trust.

Perhaps the most consequential flaw in the OSM rollout was the decision to charge for access to scanned answer scripts. Since answer books had already been digitised as part of the evaluation workflow, students were effectively being asked to pay to access and verify records that were largely their own intellectual work. This raises a fundamental question: why should students pay to inspect and verify answer scripts that had already been created and stored in digital form?

The fee-based access model also created avoidable technological and administrative problems. Students repeatedly encountered server overload, portal glitches, payment gateway failures, delayed script delivery, and difficulties accessing scanned answer books within limited time windows. Because access was restricted and linked to payment, large numbers of students attempted to download their scripts simultaneously, creating concentrated demand and congestion on the system.

More importantly, the approach missed an opportunity to build transparency directly into the evaluation process. Since answer scripts were already digitised before evaluation, CBSE could have securely shared encrypted digital copies with all students and parents by default through authenticated portals, email, or registered mobile applications. Students could then verify within a defined period that the scripts were complete, readable, correctly indexed, and genuinely theirs, including all supplementary sheets. Such a verification layer would likely have detected many issues at an early stage, including blurred scans, missing pages, wrongly uploaded scripts, and incomplete digitisation. It would also have distributed server load more evenly, rather than concentrating access requests within a paid application window. In effect, a transparency-driven model could have reduced disputes, minimised technological failures, improved public confidence, and transformed millions of students into the first line of verification for the digital evaluation ecosystem.

Proposed National Digital Evaluation Model (NDEM)

The National Digital Evaluation Model (NDEM) can build upon the strengths of CBSE’s OSM while addressing the weaknesses exposed during its rollout.

Digitisation: Under NDEM, answer scripts should first be scanned using standardised high-resolution infrastructure and securely archived with appropriate cybersecurity safeguards. Before evaluation begins, encrypted digital copies of answer scripts should be delivered free of cost to students and parents through authenticated portals, email, or registered mobile applications. Students should be given a limited verification window to confirm that the scripts are complete, readable, correctly indexed, and genuinely their own, including all supplementary sheets.

Digital Evaluation: Simultaneously, detailed subject-wise, step-wise marking schemes should be finalised and published. Once script verification is completed, evaluation should proceed through trained examiners using standardised digital platforms. Digital audit trails, centralised moderation, automated totalling checks, anomaly-detection systems, and random quality audits should continuously monitor consistency and compliance with prescribed SOPs.

Self-Verification: After evaluation, students would already possess both their answer scripts and the approved marking schemes. This would enable them to independently estimate their likely scores and compare them with officially awarded marks. Significant deviations could then trigger targeted review mechanisms, thereby reducing the need for large-scale re-evaluation requests.

Review: The model retains re-checking and re-evaluation as final safeguards but shifts the emphasis from post-result dispute resolution to pre-evaluation verification and transparency.

By combining secure technology, standardised evaluation, student verification, and digital accountability, NDEM can create a transparent, auditable, and trustworthy evaluation ecosystem for Board, university, admission, and recruitment examinations across India.

The success of MCQ-based examinations has shown that transparency, verification, and digital auditability strengthen public trust and system integrity. Descriptive evaluation must now evolve on similar principles.

The proposed National Digital Evaluation Model (NDEM) should be supported by robust SOPs, reliable scanning infrastructure, secure archival systems, high-capacity servers, structured examiner training, standardised step-wise marking schemes, strong cybersecurity safeguards, and automated audit mechanisms. Most importantly, scanned answer scripts should be securely provided free of cost to all students and parents, enabling verification of completeness, readability, and authenticity before evaluation begins.

By extending the transparency framework of MCQ-based examinations to descriptive assessments, NDEM can transform evaluation into a transparent, verifiable, auditable, and student-centric ecosystem. Importantly, the framework can be implemented immediately, helping address many issues exposed during the CBSE-OSM rollout. If adopted carefully, NDEM can become a national benchmark for digital evaluation and restore confidence in large-scale examination systems.

The writer is a Tech Education Policy Consultant, a former Professor of Computer Science at IIT Kharagpur, IIT Kanpur, BITS Pilani, and JNU, and a former scientist at DRDO and DST; Views presented are personal.