Body cameras, facial recognition, analytics: Govt institutions tap AI to fight exam cheating

Imagine this scenario playing out in a college in India in the not-too-distant future. A 19-year-old health sciences student waits in line to get into an exam hall to take his pharmacy exam. He soon realises the experience is akin to going through the security check at an airport to board an aircraft. As his turn comes up, a staff member asks the student to face a camera, which scans his face and compares it with a recent image he had been asked to submit weeks earlier.

The image is a match and the entry gate flashes a green signal, allowing him to enter. A few students who had gone before him had triggered a red signal and been asked to step aside for further scrutiny.

As the student enters the exam hall and takes his seat, he notices multiple cameras across the room. To his surprise, the proctor steps out immediately after distributing the question paper and answer sheets.

Minutes into the exam, the cameras begin moving randomly, focusing on particular students. A voice notification from a speaker inside the room reminds examinees that the centre is equipped with devices that can detect audio inputs, recording any conversation among them.

This is how the Bengaluru-based Rajiv Gandhi University of Health Sciences (RGUHS) is planning to use technology to curb malpractices, drastically reducing human intervention in the invigilation process.

Across India, there is a high incidence of cheating during examinations. From the use of chits and Bluetooth devices to individuals impersonating others and taking exams on their behalf, malpractice takes various forms. The shortage of faculty and the absence of external invigilators is a significant challenge in the effort to conduct free and fair examinations.

When the system fails, the quality of assessments is affected and cheaters get an unfair advantage over those who have worked hard on their academic performance. In an effort to combat the menace, some state-run institutions have begun harnessing technology.

RGUHS, which was established by the Karnataka government in 1996 and oversees 1,242 institutions across 31 districts in Karnataka, recently invited a tender for the “web streaming of theory examination centres” with an estimated cost of Rs 7 crore.

“We plan to establish a centralised examination vigilance and remote squad centre at the University to hold secure and malpractice-free examinations. It will have a provision to connect to the surveillance systems in all affiliated colleges,” MK Ramesh, vice-chancellor of RGUHS, told Moneycontrol.

“At present, we have thirty 50-inch TV monitors at the University headquarters to oversee 1,500 examination rooms where students from across Karnataka take their exams. However, it is challenging for our staff to physically monitor multiple centres during the approximately three-hour-long exam durations," said Ramesh.

Equipped with the database of student images along with face-recognition technology, the system will flag any impersonation attempts and unusual movements or reactions of students inside the exam hall, he added.

The AI-based analytics provide features such as intrusion detection, face detection and recognition, people counting, loitering detection, dwell time analysis, and crowd formation assessment.

“This marks the first time that a university in India is employing AI on such a large scale to combat exam malpractices,” said Ramesh. The university has around 3 lakh students enrolled in various courses, including medical, dental, AYUSH, nursing, physiotherapy, pharmacy, and allied health sciences.

“We plan to implement an audio detection feature to prevent oral dictation during the examination. This is aimed at ensuring a foolproof system for those taking medical and health science exams,” said Ramesh.

The university is not alone.

KPSC uses body cameras

The Karnataka Public Service Commission (KPSC) is also using technology to combat exam malpractices. “We have provided 3,037 body cameras for all invigilators to monitor students during examinations. We already have CCTV cameras, and the exams are also being video-recorded,” said KS Latha Kumari, KPSC Secretary, told Moneycontrol.

She added that jammers had been installed in exam halls to prevent students from using mobile phones and other electronic devices during exams.

“We also conduct frisking of all students. One hour before the exams, all invigilators use face-recognition software on their iPads to match the candidates’ images with their hall tickets/documents to prevent impersonation. In future, we may use face recognition cameras to identify those involved in malpractices,” added Kumari.

In 2020, the Tamil Nadu State medical university launched an AI-driven system, which can monitor examination venues real-time and prompt officials about “suspected” malpractices. Similarly, in February this year, the Nagpur Zilla Parishad created a district control room, leveraging technology to create an added layer of security for its state board exams.

Also, read: Students love online exams but can scores be taken seriously?

Consent and freedom

While such efforts may enhance the integrity of exams, there are concerns over the right to privacy and the sanctity of student data.

“We are not opposing the use of cameras or technology to curb exam malpractices, but they should not be used to restrict students’ freedom,” said Mahantesh B, state office secretary of the All India Democratic Students Organisation (Karnataka).

According to Prateek Waghre, policy director, Internet Freedom Foundation, there is a need for more clarity on whether these requirements for students’ personal data is being mandatorily imposed.

“It boils down to how aware students are about their data being used. Did the institution ask for consent and could it be freely exercised,” asked Waghre. He added that institutions need to inform students on how their data is being stored, used, and shared later, after the examinations.

Pallavi Bedi, senior policy officer, Centre for Internet and Society, agrees. “The important thing is: do students know how long this data will remain with the institution and can they ask for its deletion after the examination,” said Bedi.

All this comes at a time when the government is set to release rules for the implementation of the Digital Personal Data Protection (DPDP) Act, 2023, which was passed in August.

Section 17(1)(c) of the DPDP law exempts the requirements of notice and consent, among others, for the purposes of processing for “prevention, detection, investigation or prosecution of any offence or contravention of any law.”

Bedi said that this provision has not been notified yet. However, she added, the government might exempt educational and government bodies from the Act with the help of this provision, to enable them to process student data in the effort to curb exam malpractices, as cheating could be considered a breach of public order and a serious offence. The jury is still out on this.

Effectiveness of the tech

Waghre believes that there is limited proof to indicate that this medium is more effective to curb exam malpractices than the earlier method of manual invigilation.

“There can still be mismatches and cases of false positives—absolute reliability cannot be placed on this technology,” said Waghre.

Bedi said there needs to be more clarity on what tech is being used and how effective it is. She believes there could be cases where AI might incorrectly identify students for malpractices based on a slight variation in their appearances. “Being accused of cheating when they were not guilty, could end up creating mental agony for students at a time when they are already under pressure,” added Bedi.

Also, read: How a Noida-based exam security solutions firm is using AI tools to manage crowds at the Kumbh

As institutions turn to facial recognition and AI-based surveillance to bolster exam invigilation, they will have to strike a fine balance between ensuring academic integrity and safeguarding students’ rights and well-being.