New Delhi: As many as 9,606 SIM cards sold in Kerala were blocked as part of the Artificial Intelligence-based fake SIM card hunt launched by the Centre in the past two years.

As per estimates released by the Telecom department, a total of 36.61 lakh SIM cards have been blocked in the country since 2022 this way. These SIM cards were found to have been used for committing cyber fraud.

As part of the hunt, several SIM cards procured by a single person by furnishing fake details and documents have been blocked using the AI-based system called ASTR. For that purpose, as many as 87 crore SIM cards have been subjected to verification using the software. 

The 9,606 SIM cards blocked in Kerala were part of the 11,462 ones that were reported for suspicious activities. The system subjected a total of 3.56 crore SIM cards from the state for verification. Seven points of sale, which sold fake SIM cards, have been blacklisted.

The number of fake SIM cards in Kerala, meanwhile, is much lower when compared to other states.

Commenting on the operation, Telecom Minister Ashwini Vaishnav said the operation unearthed an instance in which a single person procured as many as 6,800 SIM cards in different names by providing fake details. The number of SIM cards procured by another fraudster, meanwhile, stood at 5,300.

Naveen Jakhar, Additional Director General of the Telecom department and one of the key architects of the ASTR system told Manorama that the system envisaged to prevent fake SIM cards before they were being used for plotting online frauds.

States that lead in fake SIM cards
1. West Bengal- 12.34 lakh
2. Haryana -- 5.24 lakh
3. Bihar– Jharkhand – 3.27 lakh
4. Madhya Pradesh -- 2.28 lakh
5. UP – 2.04 lakh
6. Gujarat – 1.29 lakhs

What is ASTR?
The ASTR operates by verifying the photographs provided to telecom companies by customers for purchasing SIM cards. After verifying these pictures using an AI-based system, photographs that are found to bear similarity with each other will be listed together. To enable this listing, the pictures should bear a similarity of at least 97.5 %. 

The system is capable of identifying similar faces from one crore photographs in just ten seconds. Upon zeroing in on the pictures, the KYC details will be verified to find out the differences. In most cases, the details will be different from each other and they will be cancelled after scrutiny.

WhatsApp, on its part, has expressed its willingness to remove the chat accounts issued against such numbers.