Election Commission's AI system for West Bengal voter roll revision lacks transparency
India's Election Commission plans to use an AI application during a Special Intensive Revision of West Bengal's electoral rolls to identify duplicate and fraudulent voter registrations, but the system's developer, functionality and bias testing remain undisclosed, prompting questions from opposition politicians about lack of transparency.
What happened
The Election Commission of India announced it will use an AI-based verification system during the Special Intensive Revision (SIR) exercise of electoral rolls in West Bengal, analysing facial similarities in photographs to identify individuals registered at multiple locations and prevent fake or deceased voter inclusion.
TMC Rajya Sabha MP Saket Gokhale questioned the lack of transparency, noting zero details are publicly available about who developed the application, its functionality, or whether it was audited for AI bias.
Reported impact
West Bengal Chief Minister Mamata Banerjee raised concerns about the hiring of 1,000 data-entry operators and 50 software developers from outside for the revision exercise.
Index analysis
Gokhale questioned why AI is necessary for duplicate detection when existing PDF software can accomplish the task, and raised concerns about potential political bias given the Election Commission's prior hiring of an agency associated with the BJP IT cell in Maharashtra in 2019.
Governance controls
The Election Commission has not disclosed details about the AI system's vendor, developer, or whether the application underwent bias auditing.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.