ChatGPT Has Sobering Implications for Traditional Models of Teaching and Assessment
In a recent article on the pienews.com, Nicholas Cuthbert argued that language AI tools like ChatGPT have “sobering implications for traditional models of teaching and assessment” and may compromise the academic integrity of assessments.
The Password team and its online proctoring partner, Examity, are following this debate closely, as cheating on assessments has evolved with the pandemic-driven acceleration of online testing.
Traditional Paper-based Exams
While traditional paper-based exams have not disappeared, they are not invulnerable to technology-based cheating with platforms like JasperAI, ChatGPT, or the new Bing Search. Students only need to sneak a web enabled device into the exam hall. Responsible online test providers are operating with the same rigorous procedures as paper-based exams, with even higher ratios of invigilators to students.
For example, the Password Skills Plus test has a ratio of one invigilator to four students, making it easier to catch cheating attempts during the examination and even afterwards when exams are reviewed. In addition, proctored online exams have security features such as human and AI monitoring and time-stamped tracking video, making it incredibly hard to cheat and get away with it.
Limitations Where Language Proficiency Assessment is Concerned
ChatGPT may have particular limitations where language proficiency assessment is concerned, as it is unable to imitate second language learners with mother-tongue interference, which could make its answers seem inauthentic, and possibly easier to detect.