Special Intensive Revision: AI detects 6.41 lakh anomalies in Bengaluru
technology ai | Photo credit: Suchat longthara
For the first time since the enumeration phase began, the Greater Bengaluru Authority (GBA) has revealed that there are at least 6.41 lakh “logical discrepancies” or anomalies in its jurisdiction. As of Saturday 4 p.m., only 38.81% of the enumeration forms had been digitized. The total also includes 2.66 lakh anomalies flagged in Bengaluru Urban.
Total Voters 1,03,88,363 No Mapping 22,25,260 Anomalies 6,41,766
However, the nature of these discrepancies was not disclosed in the official progress report.
GBA officials confirmed to The Hindu that artificial intelligence (AI) is being used to identify these irregularities. “An AI software tool is being used to identify irregularities. Once the BLO digitizes and submits the enumeration forms, the system will flag all the anomalies,” said a senior official.
The official added that the same software is also used to identify duplicate voters through the Demographic Similar Entries (DSE) tool, which identifies voters sharing identical demographics, such as names or fathers’ names, across electoral databases.
“The DSE tool also flags logical discrepancies resulting from name changes or other demographic details,” the official said.
However, while the GBA daily progress report lists the number of anomalies separately, the national daily progress report does not mention this category even now.
The disclosure comes even after Chief Executive Officer (CEO) V. Anbu Kumar reiterated that the anomalies will be known only after the draft electoral rolls are published.
GBA has a total electorate of over 1.03 crore. In addition to the anomalies, more than 15.83 lakh voters were categorized as “not mapped or no mapping” – indicating that they had submitted their enumeration forms without filling in the details of the 2002 electoral rolls because they were unable to trace them.
“Many voters have at least one discrepancy. Some have two or three discrepancies, such as spelling variations in documents or expansion of initials,” the official said. “But the AI also reports anomalies that are impossible. For example, it reports a discrepancy that says the gap between two offspring linked to the same parent in the previous SIR is less than nine months.”
Officials said it was difficult to estimate how many irregularities would eventually be discovered before the draft electoral rolls were published. Before the designs are finalised, ‘data shredding’ – as seen in West Bengal – will be done to remove discrepancies that may have been wrongly flagged by AI. After that, only the mechanical inspection of the remaining cases will be carried out.
Published – 26 Jul 2026 07:00 IST