SIR update: ECI to use AI-powered facial recognition to weed out duplicate voters

New Delhi: In a significant step toward strengthening the accuracy and transparency of India’s electoral rolls, the Election Commission of India (ECI) is preparing to deploy Artificial Intelligence (AI) tools to identify duplicate voters during the ongoing Special Intensive Revision (SIR) of electoral rolls.

Sources said on Friday that the advanced system will be used in the second phase of the nationwide exercise, which is currently underway in nine states and three Union Territories.

According to senior officials familiar with the matter, the AI module will utilise facial recognition technology to detect voters who may have been enrolled more than once using the same photograph.

By comparing faces across the massive electoral database, the software will flag potential cases of duplication for further verification.

“The AI-based software, through the face recognition feature, will identify voters and de-duplicate them,” a source said, emphasising that the system is meant to act as an analytical aid rather than an independent verifier.

The Commission has been increasingly concerned about reports of duplicate photographs being used during voter enrolment—particularly in cases involving migrant workers, where photographs have allegedly been repurposed to register fake or even deceased voters.

The new AI-driven approach is expected to help the ECI detect such instances faster and with greater accuracy.

However, officials involved in the exercise clarified that the introduction of AI will not replace human verification, which continues to be the backbone of the SIR process.

Booth Level Officers (BLOs) have been tasked with door-to-door verification to confirm the identity, address, and photograph of every elector.

“AI tools alone cannot ensure full transparency,” sources stressed, noting that BLOs have been instructed to physically verify each elector before finalising the information in the rolls.

The second phase of the SIR—comprising enumeration form distribution, collection, and digitisation—is expected to be completed by November 25–26, ahead of the publication of the draft electoral roll. The ECI has made rapid progress in the data-collection stage.

According to officials, Commission personnel have already distributed 50.40 crore enumeration forms, which accounts for nearly 99 percent of the target across participating states and Union Territories. Of these, more than 13.50 crore forms, or about 27 percent, have already been digitised and entered into the system.

The use of AI in electoral roll management marks a major technological shift for the Commission, which has historically relied on manual cross-checks to clean voter lists.

The initiative comes at a crucial time, with multiple state elections and the preparation cycle for the next General Election already in motion.

While concerns over privacy and algorithmic accuracy continue to be part of broader discussions around AI deployment in governance, officials maintain that the current system has been designed with strict oversight and is meant only to support—not override—human scrutiny.

The ECI is expected to share further details on AI integration once the second phase of SIR concludes and the draft rolls are prepared for public review.