In India, the Union Public Service Commission (UPSC) is all set to improve the examination credibility to prevent cheating and impersonation through AI-based CCTV surveillance and facial recognition systems. This is said to seek to improve security and integrity in the process of examination. Main findings: This section should present the main findings of the study.
It started with the UPSC recently inviting bids from seasoned public sector undertakings for two important technological solutions. Some of these are Aadhaar-based fingerprint authentication with facial recognition and the live AI/ML-enabled CCTV camera surveillance system.
To ensure security during the exams, the UPSC intends to acquire and deploy sophisticated CCTVs in examination halls nationwide to monitor candidates and other persons present during the exercise. This is done, they say, to maintain the credibility of the examination.
For each of the examination venues, at least one high-quality CCTV camera per classroom will be installed to cover all candidates who sit in a classroom of not more than 24 candidates. There will also be cameras at the entrance and exit points. CCTV will be installed in other control rooms where documents regarded as sensitive are processed.
The AI surveillance system is intended for creating real-time alarms for different situations, such as movements at the entry/exit gates if the classroom furniture is rearranged or if the camera is faulty. It will also track class activity both before and after an exam and immediately report any suspected cheating or an inactive invigilator.
This decision comes against the backdrop of scandals that surround trainee IAS officer Puta Khedkar who was alleged to have fraudulently contested the examinations and provided false identification information. Further probe is still in process to determine the truth of her stories related to disability and the OBC category. The new surveillance measures are intended to avoid the same and guarantee an impartial testing environment for all the candidates.
