English abstract
This study aimed to examine the potential integration of generative AI in the
mathematics curriculum as a supplementary tool for higher education students at the
College of Science, Sultan Qaboos University (SQU). The study also aimed to find
appropriate ways to use customized assessments to prevent academic dishonesty related
to the use of AI tools like ChatGPT. A quasi-experimental design was employed using a
convenience sample of 54 students. Mixed methods, including quantitative and qualitative
approaches, were used to gather the study's data. Quantitative methods included in-class
assessments (two exams and two assignments), while qualitative methods included
observations and semi-structured interviews. The independent variable in this study
was the use of ChatGPT by the experimental group in learning mathematics
compared to the control group, which did not utilize ChatGPT, and the dependent
variables included student performance. The quantitative analysis demonstrated a
notable difference in performance between students who utilized ChatGPT and
those who did not. Specifically, the experimental group outperformed the control
group in general assessments, whereas the control group performed better in
customized assessments. Moreover, the qualitative analysis of semi-structured
interviews indicated that participants perceive ChatGPT as a valuable tool for
enhancing their proficiency in basic mathematics. They expect ChatGPT's
problem-solving ability to improve significantly when students accurately enter
mathematical symbols and provide clear directions. Additionally, through
observation, the researcher identified the obstacles that ChatGPT encountered
when faced with particular mathematical problems, providing insights for
educators on how to modify questions that generative AI tools like ChatGPT can
solve into customized ones they cannot. This strategy could help deter academic
dishonesty. However, the study acknowledges the need for further research on the
long-term implications of integrating generative AI in mathematics education,
focusing on advanced topics. Based on the findings, some educational implications
and recommendations in this area are suggested.
