AI + Facial recognition app: Tool compares facial features to identify similar visages with high level of accuracy; experts worry about potential misuse of tech
The tool identifies faces on the basis of their structure; (left) a visual depicting how ASTR detects multiple SIM-related documents that have the same photo
AI + Facial recognition app technology has emerged as a game-changer in combating SIM card-related fraud, which has been linked to serious criminal activities, including terrorism.
ADVERTISEMENT
The Department of Telecommunications (DoT), using ASTR—an artificial intelligence and facial recognition-powered tool for SIM subscriber verification—discovered that over 32,000 fraudulent SIM cards were active in Mumbai. Subsequently, the department filed an FIR with the Mumbai Police, leading to the blockage of the cards.
Despite its efficacy, experts caution that the high likelihood of errors in this technology raises concerns about its potential misuse, potentially jeopardising human rights and civil freedoms and government misuse.
The DoT convened an ‘Awareness of Point of Sale’ workshop at the BSNL office in Juhu on Wednesday, summoning Mumbai-based nodal officers from Airtel, Vodafone Idea, Jio and MTNL, alongside retailers. During the meeting, DoT announced plans to assign a unique code to retailers for the sale of SIM cards, aiming to identify any retailer involved in selling SIM cards after receiving fraudulent documents. The measure intends to hold retailers accountable and prevent the SIM providers concerned from facing legal consequences due to fraudulent activities. “At present, retailers are involved in the fraud, causing trouble for SIM providers. A unique identification code will help in pinpointing the respective retailer in case of any fraudulent activity,” an official from DoT emphasised.
Identity chicanery
The DoT reported receiving a list of 33,484 suspected fraudulent SIM cards this year. Leveraging the ASTR system, DoT successfully detected 29,673 instances of fake documents used for purchasing SIM cards in Mumbai, leading to the disconnection of 32,024 fraudulently obtained SIM cards. Additionally, DoT filed 12 FIRs with the Mumbai police against 43 point-of-sale retailers (PoS) for selling these SIM cards after receiving counterfeit documents. They also blacklisted 105 PoS with the assistance of ASTR.
Nodal officers of telecom firms and officials attend the workshop at the BSNL office in Juhu on Wednesday
During the investigation, DoT found that a man based in Mumbai had acquired 684 SIMs under different names using forged documents but with matching photos. Similarly, another individual in a different city was found with 304 SIM cards registered under different names but using the same photo. Two more people were discovered with 378 and 210 SIM cards each. Both had used different names but identical pictures.
Discussing the effectiveness of ASTR, DoT Mumbai Director M K Jain highlighted an instance where a woman used different names before and after marriage, procuring two SIM cards using a voter ID and an Aadhar card. The woman was penalised, and her old SIM was deactivated.
How ASTR works?
ASTR operates by comparing facial features such as the shape of one’s forehead, eyes, chin and nose as well as skin colour with a very high level of accuracy. The system's proficiency enables the detection of fraudulent activities, facilitating the identification of potential offenders.
“We plan to assign a unique code to all retailers; any unverified SIM card sales will lead to their identification, the registration of FIRs and subsequent blacklisting, prohibiting them from further SIM card sales,” explained Jain.
During the KYC process for SIM cards, retailers have transitioned to the Digital Activation Process, employing E-KYC and D-KYC methods to validate customer identities. The DoT has specifically directed retailers to refrain from using flashlights while uploading residential proof such as Aadhaar cards or passports. Additionally, they are advised against using photocopies of these documents or driving licences. Retailers are required to capture clear images and solely verify the customer's photos.
On an annual basis, the DoT scrutinises 42,000 SIM card connections within the ASTR system, conducting special audits when presented with suspicious lists from agencies such as the ATS, NIA, and the Mumbai Police. The DoT has also initiated visits to the warehouses of SIM cards for further oversight.
Experts’ reservations
Leading cyber expert Yasir Arafat Shaikh, CTO and founder of Macksofy Technologies Pvt Ltd, said, “Face recognition technology powered by AI brings up a number of issues. By permitting widespread surveillance, it can violate privacy and perhaps result in the misuse of personal information. Bias and errors are a concern, as some demographic groups are more commonly mistaken than others. In addition, the technology can be utilised for monitoring and management, endangering human rights and civil freedoms.”
He further explained that the technology could also be used for tracking and control, posing threats to civil liberties and human rights.
He stated that facial recognition can facilitate mass surveillance without individuals' consent, risking privacy rights and personal data misuse. The technology can exhibit bias, particularly against women and racial minorities. Inaccuracies result in false positives, potentially categorising innocent individuals as criminals. Lastly, governments and private entities may misuse facial recognition for tracking and monitoring, which can stifle dissent and violate civil liberties and human rights, according to Shaikh.
Retailers raise concerns
During the meeting, a Khar-based retailer shared an incident where a 20-year-old had submitted an Aadhaar card to purchase a SIM card, leading to subsequent inquiries from the Nirmal Nagar police. The DoT advised the retailer not to accept an Aadhaar card if it does not correspond to the individual presenting it.
In this case, the card had been issued ten years prior, and the SIM card was obtained based on this outdated document. The DoT suggested that the retailer inform the buyer to update their Aadhaar card and only accept the updated version with the latest photo.
Additionally, the DoT instructed retailers to capture the buyer's photo in front of a white background, emphasising that photos with masks should not be accepted. It was stressed that the Aadhaar cards should clearly display the photo, address and name and possess complete details such as street names, city or house numbers. Retailers were advised to inform the nearest police station if they suspected any fraudulent activity.
32,024
No. of SIM cards blocked recently
42K
SIM connections scrutinised via ASTR every year
Subscribe today by clicking the link and stay updated with the latest news!" Click here!
