India using face biometrics to fight fraud in voting and mobile registration
Indian officials across India are increasingly using facial recognition to identify fraud in elections and SIM card registrations.
The State Election Commission (SEC), for example, deployed self-named face biometrics software by Chennai startup FaceTagr for Panchayat (village council) by-elections May 25, according to the Times of India.
According to the Times, FaceTagr‘s systems helped government workers prevent about 10,000 incidents of voter fraud.
Seemingly satisfied with those results, commissioners used the face-matching system during the June 9 urban local body by-elections, too, the Times has reported. It is not known what if any fraud was prevented.
FaceTagr compares a person’s face with their ID card in three seconds, which is “much less than the time taken by officials to verify in person,” FaceTagr CEO Vijay Gnanadesikan reportedly told the newspaper. “Overriding decisions have to be noted down and that scares away people from indulging in such acts blatantly.”
Fighting SIM card fraud in Mumbai
Outside the voting booth, Mumbai’s police and telecom departments have used facial recognition software to identify and disconnect 30,000 fraudulent SIM cards in a recent operation, according to the Hindustan Times.
The police subsequently arrested 13 people, most of whom worked in phone stores as sales staff.
“We have found that such SIM cards taken with the help of forged documents were even used by a fake call centre,” said an unnamed police officer in the case, according to Hindustan Times. Mumbai recorded a 63.7 percent increase in cybercrimes from 2021 to 2022.
The department made the catches using Telecommunications Department-created algorithm, AI and Facial Recognition-powered Solution for Telecom SIM Subscriber Verification (understandably better known as ASTR).
The Department of Telecom echoed concerns that forged documents are highly likely to be used for cyber fraud, financial fraud, and anti-national activities. The department believes that ASTR can help prevent such fraud.
The algorithm, according to the Hindustan Times, found people who allegedly were issuing fraudulent SIM cards in retail outlets. Suspects reportedly sold SIM cards using strangers’ names without their knowledge or consent to collect commissions.
They often take photos of a few people in different poses and with different appearances to pass cursory human inspection.
For example, one employee of a telecom service provider store had created 684 SIM cards in his name. The photos of 62 people were associated with 8,247 SIM cards, according to the report.