One of the biggest challenges posed to the education sector as it increasingly moves online is how to make sure students are submitting their own work and actually sitting internet-based tests themselves.
Instead of hiring a fleet of invigilators to deploy into students’ homes to peer over their shoulders, Curtin University developed the Intelligent Remote Invigilation System (IRIS) to tap into students’ devices and automatically detect instances of cheating.
Building on an earlier pilot from the Engineering Institue of Technology, IRIS is a cloud-based platform that uses machine learning and data analytics to harness the webcam and microphone on a student’s computer to monitor for coaching and check the test taker is the correct person.
Screenshots are also periodically taken to show what the students are seeing on their screens.
The potential for discomfort, extra work and privacy breaches was realised early on in the project, informing the end design to ensure students and staff weren’t unnecessarily stressed.
A cloud-based platform was settled on as the most secure, least-intrusive way to obtain data and to minimise IRIS’ footprint on students’ devices.
Furthermore, an avatar was designed to introduce students to the tool and familiarise them with how it works.
Lastly, IRIS is largely automated to reduce the workload on staff, with options for them to return to and view instances flagged by the machine learning algorithms.
Automation also allows IRIS to interact with a student to correct their behaviour as it occurs.
While this does still sound intrusive, feedback from successive pilots at Curtin shows that students are starting to view it as being similar to in-person invigilators, especially with the help of the avatar to help them through.
IRIS also adheres to the university’s strategic plan for 2017-2020 of innovating in ways that can also support commercialisation.
Curtin hopes to see IRIS exported to the rest of Australia’s 700 Vocational Education and Training providers following further testing and implementation across its own courses.
While IRIS has so far been used in six units across three faculties, expansion is underway as more staff look to try the solution for themselves.