Story image

Turnitin looks to AI to detect student cheaters

07 Oct 2019
Twitter
Facebook

Turnitin is almost an essential part of assignment submission at schools and universities in the digital age. Turnitin is essentially an online plagiarism checker to make sure that your assignment is your own work.

Now Turnitin is incorporating artificial intelligence and machine learning (AI and ML) to tackle the problem of contract cheating, which is when students hire somebody else to write their assignments and then they submit it as their own.

The problem is rife in countries like Australia, which is why Turnitin and Australian researchers from Deakin University decided to incorporate AI and ML into their research.

According to researchers Phillip Dawson, Wendy Sutherland-Smith and Mark Riksen, this is the first quantitative empirical study of its kind to find out of AI and ML can increase Turnitin’s detection rates of contract cheating.

As part of the study, 24 experienced markers used Turnitin’s Authorship Investigate tool to evaluate 20 student assignments, which included 14 genuine assignments and six that were purchased from cheating sites.

While Turnitin’s AI and ML detections weren’t perfect (it detected 59% of all cheating cases), it’s a step in the right direction for more intelligent cheating detections. Without AI and ML, markers were only able to detect 48% of cheating instances.

The AI and ML systems were able to analyse sentence complexity, sentence length, and other stylometrics, as well as document information such as date created and last modified.

While it doesn’t explicitly decide whether cheating has occurred, it crunches the statistics to recommend further investigation.

“In addition to potentially improving detection rates, authorship analysis approaches using machine learning also offer benefits in terms of raising awareness about contract cheating and efficiently providing evidence if they wish to take their suspicions further,” explains Deakin University’s Centre for Research in Assessment and Digital Learning associate director Phillip Dawson.

Turnitin is using the research to develop a better machine learning prediction model. According to Turnitin, it will help replicate the ‘gut feeling’ a marker gets when they suspect a student is cheating.

“Whilst Authorship Investigate was in early stages of development when this study was conducted, we’re pleased to see the value of the tool in the detection process, in bringing together all submissions made by a student and allowing rapid scanning of key points of evidence,” adds Turnitin principal product manager Mark Ricksen.

“Collaboration with higher education institutions and industry enables us to constantly test and iterate our tool so it can be used in a better, faster and more impactful way by markers to address contract cheating. There’s also potential for the software to speed up the investigation process by highlighting submissions of concern by a student and potentially determining the direction and focus of any investigation.”