The study explores the relationship between AI literacy and prompt engineering skills in the context of using large language models (LLMs) for educational purposes. Prompt engineering involves formulating precise and well-structured instructions to elicit desired responses from LLMs, optimizing interaction effectiveness. The research introduces a skill-based approach to prompt engineering and examines how non-experts' AI literacy affects their prompt engineering abilities. The study found that higher-quality prompt engineering skills predict better LLM output quality, suggesting that prompt engineering is a crucial skill for goal-directed use of generative AI tools. Additionally, certain aspects of AI literacy can enhance prompt engineering and the targeted adaptation of LLMs in education. The findings highlight the need for integrating AI educational content into curricula to enable students to effectively use generative AI tools like ChatGPT. The study also provides qualitative insights into students' intuitive behaviors towards LLM-based AI systems, revealing that students often exhibit human-like conversational structures and anthropomorphize AI systems, which can impact the quality of their prompts and LLM outputs.The study explores the relationship between AI literacy and prompt engineering skills in the context of using large language models (LLMs) for educational purposes. Prompt engineering involves formulating precise and well-structured instructions to elicit desired responses from LLMs, optimizing interaction effectiveness. The research introduces a skill-based approach to prompt engineering and examines how non-experts' AI literacy affects their prompt engineering abilities. The study found that higher-quality prompt engineering skills predict better LLM output quality, suggesting that prompt engineering is a crucial skill for goal-directed use of generative AI tools. Additionally, certain aspects of AI literacy can enhance prompt engineering and the targeted adaptation of LLMs in education. The findings highlight the need for integrating AI educational content into curricula to enable students to effectively use generative AI tools like ChatGPT. The study also provides qualitative insights into students' intuitive behaviors towards LLM-based AI systems, revealing that students often exhibit human-like conversational structures and anthropomorphize AI systems, which can impact the quality of their prompts and LLM outputs.