Artificial versus natural intelligence: Overcoming students' cheating likelihood with artificial intelligence tools during virtual assessment

Abstract

Assessment techniques need to evolve beyond traditional methods in light of the rapidly developing artificial intelligence (AI) tool technologies, such as Copilot, Bard, and ChatGPT. These AI-powered Chatbot is designed to appear similar to human speech or text and present information conversationally, making them tenable options for student assessment support worldwide. Consequently, to take advantage of the weaknesses in the AI system and foster a creative attitude in their pupils, educators must reconsider their approach to evaluation. The study conducts a comparative experiment on two different assessment methods—the traditional questioning strategy (Experiment I) versus the alternative or modified strategy (Experiment II), to assess how well the AI tools perform in the assessment and how the new technique can deter students from engaging in academic dishonesty. According to the study in Experiment I, the AI-Chatbot had a 100% positive response correlation, but in Experiment II, it had a shockingly low positive response correlation. Comparably, pupils who use AI-Chatbot and those who do not have significant performance disparities (α = 0.05, p-value < 0.001; 1.8331). Inferentially, AI-Chatbot helped students a lot in Experiment I but did considerably less in Experiment II. In other words, Experiment II's questioning approach outperforms the AI tools' level of competence. The study comes to the conclusion that if AI is effectively harnessed, human natural intelligence will always be able to overcome the challenges posed by these powerful AI technologies.