Assisting academics to identify computer generated writing

Abstract

Authentic writing is an important aspect in education and research. Unfortunately, academic misconduct occurs among students and researchers. Consequently, written articles undergo certain detection measures and most teaching and research institutions use a range of software to detect plagiarism. However, state-of-the-art Automatic Article Generator (AAG) writing powered by Artificial Intelligence provides a new platform for new types of serious academic misconduct that cannot be easily detected and even if they are detected, can be hard to prove. The main objective of this study is to raise awareness of these tools among academics. This paper first explains the features of AAG writing, then investigates whether academics can distinguish AAG writing from human writing and whether raising the awareness of AAG between academics can improve their ability to detect AAG writing. A case study showed how difficult it is for academics with no knowledge of AAGs to identify this writing. A survey was used to indicate how a training session can improve the ability of detecting AAG writing. The results show that raising awareness training increased the academics’ ability to detect AAG writing. Lastly, the possible solutions to mitigate the academic integrity issues associated with AAG writing have been discussed.

Ähnliche Arbeiten