Abstract:
Artificial Intelligence-Large Language Models (AI-LLMs) have emerged as powerful tools with diverse applications in educational settings. This research aims to investigate the current landscape of AI-LLMs and their utilisation in academic contexts, while identifying potential risks and challenges associated with their use. The study collected quantitative data through an online administered questionnaire, with a total of 114 lecturers, 147 students and 14 administrators participating. The study also examines existing policies, guidelines, and practices related to academic integrity and educational ethics at CUT (Chinhoyi University of Technology). Finally, strategies and recommendations are proposed to address the concerns and safeguard educational ethics in the use of AI-LLMs at CUT. The analysis of a dataset comprising responses from students and administrators at CUT reveals that AI-LLMs, such as ChatGPT and Gemini, are commonly utilised in diverse academic activities. These activities range from writing research papers and seeking information to generating ideas and checking grammar. The programmes of study where AI-LLMs find common use include STEM, business/economics, and non-technical subjects. However, the research also highlights a number of challenges associated with the use of AI-LLMs. These challenges include the lack of adequate information and human reasoning, shallow responses devoid of local applicability, inaccuracies, and incomplete information. These issues underscore the need for caution and critical thinking when utilising AI-LLMs in academic work. The research findings indicate that CUT does not currently have explicit guidelines and policies in place regarding the use of LLMs (Large Language Models). To address the concerns and safeguard educational ethics, several strategies and recommendations are proposed. These include clear communication of academic dishonesty guidelines, guidance and training for students on utilising AI-LLMs, encouraging disclosure of AI-LLM use in academic work, enhancing the accuracy and applicability of AI-LLMs, promoting contextual understanding and critical thinking, and regular evaluation and updates of policies and guidelines. In conclusion, this research sheds light on the current landscape of AI-LLMs in educational settings and the associated risks and challenges. Through examining and proposing policy-driven strategies, the research endeavors to support the responsible and ethical utilization of AI-LLMs (Large Language Models) at CUT and other comparable institutions. Further research and collaboration with stakeholders are necessary to implement comprehensive measures that address the concerns and safeguard educational ethics in the evolving landscape of AI-LLMs.
