The growing adoption of Large Language Models (LLMs) underscores the increasing need to develop effective prompting strategies to improve learning outcomes. To this end, this paper presents a Teaching Innovation Project (TIP), called Analysis and Development of Artificial Intelligence Prompts for Learning to Attack and Defend Computer Systems, which faces the challenge of applying new LLMs and chatbots, such as ChatGPT, to improve the teaching-learning processes on cybersecurity of students at the University of Cadiz (UCA), Spain. Specifically, this challenge is addressed by studying, analyzing and creating new prompts to facilitate the teaching and learning of ethical hacking and security auditing. The prompts were designed to guide students through key phases of ethical hacking, including information gathering, asset scanning, and exploitation, as well as the implementation of security measures. The TIP was conducted successfully with undergraduate students in the Security in Computer Systems course and master’s students in the Audit and Risk Analysis course. The results obtained demonstrate that integrating carefully designed prompts can significantly improve cybersecurity teaching and student academic performance.

Keywords: Cybersecurity, artificial intelligence, prompt, large language model, ChatGPT.