Adversarial learning: teaching cybersecurity through creative battles between students and chatbots

Resumen

The use of Artificial Intelligence models to facilitate some professional activities or daily living actions has been an increasing tendency for the last years. Among students has arisen the sentiment that everything they can find “through the Internet” is not worthy to be learnt. As a consequence, current college students have a very low motivation for theoretical or abstract knowledge. An increasing absenteeism is the main symptom, as well as a total lack of interest for all those competences that are not directly related to technologies and ability with a clear a direct application in companies and the labor market. In order to address this situation, a pilot experience was planned and carried out at Universidad Politécnica de Madrid. As a proposal, during year 2023/24 a league was organized. Students could participate in groups (between four and six people). The league was composed of a sequence of battles, where students must compete against chatbots and other similar tools. Students must develop a product, usually software, following a creative process. In this pilot experience, students from cybersecurity courses were selected as participants. During the battle, students could use chatbots to get some initial designs. Later, students should analyze the responses from the intelligent models and propose some improvements. Finally, they need to prove that the proposed solution improves the one from the chatbots in any dimension. Results confirmed a significant decrease in the absenteeism rate and a relevant improvement in the students’ motivation, according to the official surveys.