ADVANCING CONVERSATIONAL AI: BEST PRACTICES IN PROMPT ENGINEERING FOR ENHANCED CHATBOT PERFORMANCE

Abstract

Technologies related to artificial intelligence are progressing at a rapid pace. This evolution brings with it new possibilities and problems, one of which is the growing usage of large language models (LLMs) in human-AI interaction. One method that can be utilized to optimize goal-directed interactions with LLM-based artificial intelligence systems is known as prompt engineering. This refers to the process of generating well-structured instructions to receive the desired information or responses from an LLM program. On the other hand, there is a dearth of studies on the implications of AI literacy on prompting behavior or on the perception of LLM-based AI systems by those who are not specialists through prompt engineering. This is an especially crucial consideration when thinking about what LLMs mean for universities. Here, we take a look at this problem from every angle, provide a skill-based method to prompt engineering, and give serious thought to how students' level of AI literacy affects their ability to use prompt engineering. Additionally, there is qualitative data on the natural behaviors of students concerning AI systems that are based on LLM. For the goal-directed application of generative AI approaches, the results demonstrate that timely engineering is an essential ability. This is because it forecasts the quality of the output of LLM. In addition, the findings indicate that particular domains of AI literacy have the potential to facilitate the facilitation of speedier engineering and more individualized LLM adjustments inside the classroom setting. To create a hybrid intelligent society where students can utilize generative AI technologies like ChatGPT effectively, we, therefore, propose that the present curriculum incorporate AI instructional content.