JAIPUR: After a major recruitment exam was marred by allegations of impersonation and paper leaks, Rajasthan’s Special Operations Group turned to artificial intelligence — scanning millions of photographs to identify candidates who had appeared in place of others. The result, officials say, is a growing list of arrests and a new model for policing public examinations.
The Leak That Prompted a Technological Shift
In December 2022, the General Knowledge and Educational Psychology paper for Rajasthan’s Senior Teacher Recruitment Exam was leaked, forcing authorities to cancel the test and reconduct it in January 2023. The cancellation rattled thousands of candidates and deepened concerns over the integrity of government recruitment examinations in the state.
As the investigation unfolded, the Rajasthan Special Operations Group, or SOG, began receiving complaints that as many as 14 candidates had employed “dummy” candidates — individuals who appeared in examination halls in place of registered applicants. By that point, 12 original candidates had already been arrested. But investigators had identified only three of the stand-ins who had actually taken the test.
Tracing the remaining suspects proved difficult. According to officials, those who had impersonated candidates had filled out application forms using the genuine applicant’s personal details, uploading only their own photographs. Because the same photograph appeared on the admit card, they were able to sit for the examination without raising suspicion at test centers.
Investigators were left with little evidence beyond photographs. Even a public announcement offering a ₹10,000 reward for information failed to yield immediate results. Faced with limited leads, the SOG turned to a technological solution.
Certified Cyber Crime Investigator Course Launched by Centre for Police Technology
Building a Digital Net
Under the leadership of SOG Inspector General Sharat Kaviraj, the agency commissioned the development of a specialized software tool designed to detect impersonation in government examinations. The system was integrated with Rajasthan’s Single Sign-On (SSO) ID system and the One-Time Registration (OTR) database — platforms mandatory for candidates applying to state government posts.
Since 2018, more than 50 lakh candidates have appeared in various government examinations in Rajasthan. The SOG now has access to that extensive database, which includes detailed applicant profiles.
According to officials, the OTR profile requires candidates to provide a photograph, mobile number, Aadhaar verification, parents’ names, date of birth, email address, educational qualifications and signature. Every examination application is submitted through an SSO ID, linking applicants to this centralized data system.
Investigators began uploading photographs of suspected dummy candidates into the new software. Using artificial intelligence-based facial recognition tools, the system compared those images against the SSO and OTR databases. Within a short period, officials said, the software identified matches.
On Feb. 16, the SOG conducted coordinated raids at six locations — including Andaman, Kolkata, Jalore, Kota and Jaipur. Five reward-listed dummy candidates and one additional suspect were arrested.
Recently, the SOG reported the arrest of five more dummy candidates in connection with the 2022 recruitment exam — three MBBS students and two government teachers. Officials said the arrests were made using the newly developed software.
Closing the Loophole
The effectiveness of the software, officials said, rests in part on a change to the OTR process itself. Before 2025, the OTR system did not require a live photograph at the time of profile creation. That gap, investigators said, allowed dummy candidates to upload someone else’s image while entering the original candidate’s personal details. With the same photograph printed on the admit card, impersonators could pass identity checks at examination centers. The loophole has since been closed. Live photo verification is now mandatory, ensuring that real-time images are captured and stored in the system.
Officials noted that many dummy candidates also appear for examinations using their own SSO IDs. Once live photographs were introduced, their authentic images were recorded in the database — creating a digital trail. When those individuals later appeared as stand-ins for other candidates, the facial recognition software could detect similarities between stored images and those used in suspect applications.
The system is designed to scan lakhs of photographs and can identify potential matches even at a 70 percent similarity threshold. It is also capable of detecting changes in age, hairstyle or other aspects of appearance over time, according to officials. By matching not only photographs but also signatures and profile details across applications, investigators say they have expanded the scope of their inquiry beyond the initial complaints.
A Growing Database of Suspects
The SOG describes the effort as part of a broader data analysis initiative. Inspector General Kaviraj said the department has analyzed records of more than 50 lakh candidates so far, matching photographs and signatures to identify patterns consistent with impersonation.
According to SOG Additional Director General Vishal Bansal, a database of more than 2,000 suspected dummy candidates has been prepared. The database contains detailed background information on those individuals. So far, 145 dummy candidates have been arrested, officials said.
The department has indicated that scanning of past applications will continue, with the aim of identifying both those who appeared as impersonators and those who benefited from such arrangements. Officials have said the technology will now be used regularly to prevent fraud in future examinations.
In public statements, the SOG has warned that individuals who appear as dummy candidates for money are risking their futures. With the integration of artificial intelligence tools into the state’s recruitment infrastructure, officials maintain that offenders will eventually be identified and face legal action.
For Rajasthan’s examination system — long strained by allegations of leaks and malpractice — the shift represents a move toward centralized digital scrutiny. As the investigation continues, authorities say the software will keep scanning, comparing and flagging matches across a database that now spans millions of applicants.
