CBSE to Scale AI-Based Evaluation System for 2026 Board Exams

Teacher training for on-screen evaluation is likely to begin early next year before the rollout for 2026 Board exams.

The Central Board of Secondary Education (CBSE) will expand its AI-driven evaluation and digital checking system for the 2026 board exams. A senior official close to the development said that the move is aimed at improving accuracy and reducing errors in the evaluation process.

“CBSE has tested AI-supported tools and on-screen marking over the past exam sessions. The results have been promising, and there are plans to take it to more regional centres for the next exam cycle,” the official added speaking to Shiksha.com.

CBSE first tried out AI-based evaluation during the 2024–25 session, to study variations between theory and practical marks. The analysis revealed unusual differences in scores across hundreds of schools, leading to advisories that asked its affiliated schools to keep internal assessment practices consistent with Board norms.

This data exercise helped identify outlier cases where marks awarded internally didn’t match the students’ written performance. The exercise became CBSE’s first large-scale use of technology for moderation.

On-Screen Checking to Expand

For the upcoming examination, the Board will increase use of digital answer-sheet evaluation, where teachers will mark scanned copies instead of handling large paper bundles. The digital setup will help cross-checking easy for consistency and will make moderation quicker.

Reportedly, a digital evaluation and AI centre has already been set up at Dwarka, Delhi, where early pilots were conducted. Teacher training for on-screen evaluation is likely to begin early next year before the rollout for 2026 exams.

Human Oversight to Stay

CBSE official clarified that the system is not meant to replace human evaluation. Teachers will continue to check answers, while AI tools will be used only for data analysis, error detection, and moderation support.

The Board expecting this mix of human and machine review will make evaluation process faster and more transparent.

Read More: