Remote online assessment of knowledge has been increasing in recent years and has accelerated with the current COVID-19 pandemic situation. The ability to conduct tests virtually using technology offers excellent flexibility and provides opportunities to candidates that were impossible otherwise. However, it can also carry some problems, including fraudulent means to answer questions, leading to biased results and unfair decisions. This phenomenon may significantly impact results in exams and lead to the certification of professionals who are not qualified to practice in the labor market. Therefore, it is essential to inhibit, mitigate, and perhaps even extinguish cheating, ensuring a fair result. Advanced computing, namely machine learning, has become more accessible and provide ways to apply advanced statistical tools to detect cheating in online tests. In real-time, the automation of this process is still in its infancy but is ever more pressing to develop.
To attain this objective, we have established the following tasks:
1. A systematic review of statistical cheating detection methods (ICVS).
2. Characterization of students’ and evaluators’ perspectives concerning cheating (ICVS).
3. Module validation in a documented cheating case (ICVS).
4. Development of a computer cheating detection module (iCognitus).