CBSE OSM 2026: How India's Biggest Digital Evaluation Crisis Ended

From Ambition to Crisis

In February 2026, the Central Board of Secondary Education announced it would move all Class 12 board examinations to on-screen marking (OSM) — scanning every answer booklet and having examiners grade them digitally. The stated objectives were sound: eliminate totalling errors, accelerate result declaration, and bring India's largest board examination system in line with international standards.

By July, those objectives had been replaced by a very different set of headlines.

A 7-year low pass rate. Approximately 1.5 million cyberattack attempts on the marking portal. Thousands of students reporting blurred scanned answer sheets they could barely read. Two senior officials — Chairman Rahul Singh and Secretary Himanshu Gupta — transferred mid-cycle. A Supreme Court intervention. And a nationwide outcry that one social media thread labelled the 2026 batch "the worst-treated batch in CBSE history."

On September 25, 2026, the Appointments Committee of the Cabinet confirmed Mandeep K. Bhandari, a 2001-batch IAS officer, as the new CBSE Chairperson. His mandate, implicit in the timeline of his appointment, is to rebuild institutional credibility in digital evaluation — and to ensure the 2027 cycle does not repeat what 2026 exposed.

This post reconstructs what went wrong, why it went wrong in predictable ways, and what institutions planning their own digital evaluation programmes must now build in before they go live.

The Five Failure Points

1. Vendor Selection on Price, Not Capability

Early reporting and RTI responses indicated the OSM contract was awarded to the lowest bidder. In a system handling roughly 42 lakh answer booklets for Class 12 alone, the cost of under-specifying scanning throughput, server capacity, and security architecture is not a marginal risk — it is a near-certainty of failure at scale. The resulting platform could not sustain concurrent examiner load during peak marking windows, crashed repeatedly, and had scanning quality parameters that were not independently validated before deployment.

Any institution selecting an OSM vendor on price alone will produce the same outcome at whatever scale it operates. The procurement criteria for an examination-critical platform must weight security certification, disaster recovery capability, scan-to-examiner latency guarantees, and reference deployments at comparable scale.

2. Training That Arrived After the Examination

Teachers across Delhi government schools — organized through the Government School Teachers Association — publicly requested CBSE delay OSM implementation to the next academic session, citing that structured digital evaluation training had not been completed before answer books arrived. CBSE proceeded. The predictable result was variance in examiner technique, inconsistent annotation, and a marking pipeline where human errors compounded software errors.

Training for OSM is not analogous to training for a new software tool. It requires examiners to develop reliable judgement about scan quality before they begin, to understand how partial-answer rendering on screen differs from reading continuous handwriting on paper, and to know the escalation path when a scanned image is defective. None of this can be compressed into a single orientation session held after papers have been collected.

3. Cybersecurity as an Afterthought

A 19-year-old security researcher publicly disclosed critical vulnerabilities in the OSM portal, including pathways that could permit unauthorised access to examiner accounts and mark modification. CBSE acknowledged the vulnerabilities and described them as "contained" — a response that arrived after the disclosure was already public, not before.

Examination portals are among the highest-value targets for grade manipulation. The attack surface includes examiner authentication, answer book assignment, mark entry APIs, and the aggregation pipeline that produces final results. Each of these must pass independent penetration testing before a single answer book is uploaded. CBSE's timeline suggests this testing either did not occur or did not meet the threshold required.

4. Scan Quality Without Quality Gates

Students and parents obtained photocopies of graded answer sheets and found images with heavy shadows, cut-off text at binding edges, and pages rendered at resolutions that made handwriting illegible. This is a scanning infrastructure problem, not an examiner problem. If the digitisation pipeline does not enforce a minimum DPI standard, validate that every page has been captured within image boundaries, and flag books with low-confidence scan scores for manual re-scan before distribution to examiners, illegible images will reach the marking queue.

The answer to "what if an examiner cannot read the answer" in a well-designed OSM system is a documented escalation to a quality-review queue, not an arbitrary score or a blank mark. CBSE's 2026 system appears not to have had this gate.

5. No Grievance Infrastructure to Match Scale

When results were published and students began applying for photocopies, re-evaluation, and mark verification, the portal handling these requests failed repeatedly under load. Students reported failed payment transactions that debited amounts without generating confirmation, weeks-long delays in receiving scanned copies, and no clear timeline for redressal.

In a manual evaluation system, a re-evaluation request involves retrieving and redistributing a paper bundle — slow but tractable. In an OSM system, the same request should be substantially faster: the scanned image already exists, the original examiner's annotations are digitally logged, and a second examiner can be assigned without physical logistics. CBSE's 2026 system was unable to deliver on this inherent advantage, which amplified the perception that digital evaluation had made things worse, not better.

What the Leadership Change Signals

Mandeep K. Bhandari's appointment on September 25 comes with a clear institutional context. His predecessor was transferred in the middle of the OSM controversy. The Ministry of Education signalled, through the timing, that continuity in leadership was not acceptable after an examination cycle of this scale failed this publicly.

For the wider examination system, the CBSE experience has done two things simultaneously. It has raised the political cost of poorly executed OSM programmes, meaning future rollouts — at state boards, affiliating universities, and central universities — will face more scrutiny before deployment. It has also accelerated voluntary adoption among institutions that had been undecided, because the failure map is now explicit: institutions that do not repeat CBSE's five errors have a documented basis for confidence.

Punjab's School Education Board had already announced OSM for its 2026 matriculation cycle before the CBSE controversy peaked. Its phased approach — beginning with one subject — reflects exactly the kind of incremental de-risking that CBSE bypassed by going full-scale in year one.

What Institutions Must Build In

Universities and affiliated systems planning OSM deployments should build the following non-negotiables into their implementation plan before a single answer book is scanned:

The CBSE 2026 experience is not an argument against digital evaluation. It is an argument against inadequate preparation for digital evaluation at scale. The institutions that use this failure map carefully will implement systems that are more reliable, more auditable, and more defensible than the physical evaluation pipelines they replace.

---

Related Reading