With all classes in going online, many students are now slowly coming to grips with home-based learning. While many university students seem to take this change well, many have complained about taking their exams online as well.
A common complaint involves cheating – since students are taking exams from the comfort of their homes, it’s a fair grievance.
Using artificial intelligence to deter cheating
At least two local universities are trying to tackle this issue. Both Singapore Management University (SMU) and Singapore Institute of Technology (SIT) are using using artificial intelligence to deter cheating in online exams.
The first step is in ‘locking’ the students’ browsers so that they won’t be able to access other websites to capture screenshots or texts until the exam is over.
Students will then have to take short videos of their locations, such as the room, under the study table, or even inside pockets, before pointing the webcam to record themselves as they take the exam. This is when an artificial intelligence takes over, tracking their eye movements to determine where and what they’re looking at.
As the AI algorithm identifies movements or the presence of someone new, the video recordings are reviewed by the course instructor. The tool seemed effective to deter cheating – instructors can review the video recording of every student in class individually, unlike in a classroom setting where it may be difficult to do so.
While the tool is not a guarantee against cheating, SMU has been running online exams using Cisco’s webex interface for the past week to conduct closed-book exams. However, some SMU students didn’t need to use webcams – they had take-home exams which were open-book, and students were given anywhere from 3 hours to 3 days to finish, with the answers run through a plagiarism checker.
Online proctored exams challenges student privacy
In Australia, universities like the Australian National University, University of Queensland, and University of Sydney are using software like ProctorU and Proctorio to monitor students during in-home exams.
While both platforms also use artificial intelligence, including machine learning and facial detection technology, to verify students’ identities and spot suspicious activity during exams, they also require the students’ biometric information (ie. photo ID), with potential access to their computer’s webcam, microphone, and keystrokes.
However, privacy experts criticise this plan, citing that these software may be too invasive. Universities responded by ensuring that only authorised university staff would have access to the data collected by the programmes.
Medical students take final open-book exams online
For the first time ever, Imperial College London put 280 of its final-year medical students through two online exams from home last month. This is the first digital ‘open book’ exam delivered remotely for the university, meaning that students can access any resource material they may need – online or otherwise.
The exams present a new way to assess the students, testing their ability to diagnose a patient’s condition. Students were presented with a patient and given their history, clinical findings, and data like blood tests – and given 150 questions to be answered in 3 hours, or 72 seconds per answer.
Even if they could access the internet, the answer would not be readily available, as it mimics a real-life scenario of diagnosing a patient. To ensure students weren’t exchanging notes, the order of questions were randomised for each student.
An online open-book exam may be a new era for medical assessment – or any exam that requires critical assessment of course material, like law or social science.