Odisha Plans AI-Powered Exam System and 52-KPI State Ranking Framework

What Happened in Bhubaneswar

In September 2026, officials from Odisha's Department of Higher Education (DHE) presented an ambitious digitalisation roadmap at a two-day Vice-Chancellors' Conference held at Lok Bhavan, Bhubaneswar. Chief Minister Mohan Charan Majhi inaugurated the event alongside Governor Hari Babu Kambhampati.

The proposal that drew the most attention: a state-wide AI-powered examination system that automates question paper management, monitors examinations in real time, and — for the first time at state scale — automates the grading of answer scripts. Parallel to this, the state is rolling out a 52-KPI institutional performance framework that will rank every public university across four pillars: students, faculty, infrastructure, and activities.

This was not a study group recommendation. It was a policy announcement, with MoUs already signed on the same day with CSR Box, Nandi Foundation, and OKCL. Odisha has moved quickly in 2026: the state's higher education sector enrolled over 7.5 lakh students in government universities as of 2025-26, and the DHE has been under pressure to modernise evaluation since repeated result delays and a high-profile paper leak in 2025.

What the AI Examination System Will Do

The DHE proposal covers the full examination lifecycle.

Question Paper Management

Question paper leaks remain a structural weakness in Odisha's examination system — the state saw a paper leak at MKCG Medical College's PG entrance in 2026, compromising the credibility of the process. The AI system will manage question bank access, enforce time-locked delivery, and create tamper-evident audit trails for every question pulled from the bank.

The question bank itself will be built in alignment with NEP 2020 and the National Credit Framework, ensuring questions are mapped to learning outcomes and competency levels rather than being drawn from arbitrary repositories.

AI-Assisted Answer Script Evaluation

The most consequential part of the announcement is AI-assisted grading of answer scripts — reducing human intervention in the marking process for select subject types.

AI grading works well for structured, bounded responses: short answers, fill-in-the-blanks, and numerical problems. For extended responses and essays, the technology functions best as a first-pass scoring assistant — flagging answers for human review when scores fall outside predicted ranges, detecting copied responses, and ensuring that every answer is at least seen by an evaluator. Odisha's proposal appears to use this hybrid model rather than fully autonomous grading, which is consistent with best international practice.

This is not the same as CBT (computer-based testing). Students will still write on paper. The difference is in the back end: answer sheets will be scanned, and AI tools will support evaluators rather than replace them.

Live Examination Monitoring

The system will monitor examinations in real time, tracking invigilator attendance, flagging unusual patterns, and maintaining a digital record of each examination session. This addresses a longstanding problem in centre-based examinations: discrepancies between declared centre conditions and what actually happens on the ground.

The 52-KPI State Ranking Framework

The second major announcement — the state ranking framework — is a significant departure from how Indian state governments have traditionally assessed their public universities.

Odisha will evaluate every public university across 52 key performance indicators grouped into four pillars:

The framework draws directly from NAAC's revised criteria and NIRF's methodology — with a deliberately state-contextualised focus that national frameworks miss. A state technical university in Koraput faces different constraints than an autonomous college in Bhubaneswar. The state ranking framework allows Odisha to track improvement trajectories even for institutions that are years away from competitive NIRF ranking positions.

Crucially, examination-related KPIs appear across multiple pillars. Result declaration timelines, revaluation request rates, student grievance resolution, and examination infrastructure quality all factor into the scoring.

Why This Matters Beyond Odisha

Odisha is not alone in this direction, but it is one of the few states to announce a comprehensive, officially endorsed roadmap at this level of specificity. The announcement is significant for at least three reasons.

It Signals Scalability of AI-Assisted Grading

The debate in Indian higher education has moved from "should we digitise?" to "how much should AI do?" Odisha's announcement — by a state with 7.5 lakh students and limited per-institution resources — suggests that AI-assisted grading is now seen as a credible tool for institutions that cannot afford the full infrastructure of a mature digital evaluation setup.

It Links Examination Quality to Institutional Rankings

By building examination KPIs into a state ranking framework, Odisha is making a structural connection that national frameworks often leave implicit. An institution that takes 120 days to publish results, generates a high revaluation backlog, and has unresolved student grievances about evaluation will score lower in the state framework — creating a direct accountability mechanism.

It Reflects the Post-CBSE OSM Correction

The CBSE OSM crisis of 2026 — where a rushed, full-scale deployment of on-screen marking triggered 4 lakh revaluation requests and a national controversy — created a cautionary template. Odisha's phased, AI-assisted approach (rather than full OSM rollout at once) reflects a deliberate choice to avoid the same trap. The question bank portal, the monitoring layer, and the hybrid AI grading model together suggest a system designed for controlled rollout rather than overnight transformation.

What Universities in Other States Can Learn

The Odisha model offers three practical takeaways for universities and state governments evaluating their own exam modernisation roadmaps.

Start with the question bank. The most immediate risk in Indian examination systems is not evaluation quality — it is paper leaks and question integrity. A secure, AI-governed question bank with competency-mapped questions and controlled access is the highest-leverage first investment.

Separate monitoring from evaluation. Live examination monitoring and AI-assisted evaluation are distinct functions. A state can implement the former within a single academic cycle using existing infrastructure. The latter requires platform procurement, evaluator training, and answer sheet scanning. Doing both simultaneously is avoidable risk.

Use a state ranking framework to create accountability upstream. When examination quality factors into a visible state ranking, university vice-chancellors and examination controllers have an incentive to improve that national NAAC or NIRF visits alone do not generate between accreditation cycles. Odisha's 52-KPI model is a structural accountability mechanism, not just a data collection exercise.

What Comes Next

The DHE has not published an implementation timeline for the AI examination system. The MoUs signed at the conference cover skill development and digital literacy rather than examination platform procurement — suggesting that the examination technology rollout is in an earlier planning phase.

The 52-KPI ranking framework, however, is further along: the conference materials reference a rollout of performance monitoring starting in the 2026-27 academic year.

For the full-scale AI examination system, the realistic expectation is a pilot in select universities during the 2027-28 academic year, with broader rollout contingent on the quality of the pilot outcomes.

Related Reading