2026 ASEE Annual Conference & Exposition

Artificial intelligence (AI) and cybersecurity are converging rapidly, reshaping
how digital threats are analyzed and mitigated. Despite this shift, cybersecurity
education has been slower to incorporate hands-on learning experiences that
integrate machine learning, data science, and security in an applied manner.
This paper presents \textbf{GenAI-PassLab}, a modular, research-informed
laboratory designed to support password security education through generative
AI. The lab enables students to engage directly with a generative language model
(GPT-2) for password generation, strength evaluation, and adversarial analysis,
bridging recent research advances with instructional practice.

GenAI-PassLab addresses a persistent gap in computing and information technology
curricula: the lack of accessible laboratory infrastructure for teaching
\textit{AI/ML/DS for Cybersecurity} (AMD4CYB). While established platforms such as
SEED Labs and Labtainers have advanced hands-on cybersecurity education, few
resources integrate generative AI techniques for exploring authentication and
password security. GenAI-PassLab extends this landscape by allowing students to
examine how generative models learn password patterns, assess training data bias,
and compare heuristic-based and AI-driven password evaluation approaches.

The lab is implemented using the Labtainers platform, providing portability and
reproducibility through containerization. Students interact with a
self-contained environment that includes custom GPT-2 models trained on
real-world password datasets such as RockYou and Have I Been Pwned (HIBP). Guided
activities emphasize model behavior, adversarial reasoning, and ethical
considerations related to the use of leaked credential data. Supporting
materials align with ABET Computing Accreditation Criteria, ACM CSEC 2017
Knowledge Areas, and NIST NICE competencies.

Initial classroom deployments in upper-level undergraduate computing courses
indicate high student engagement and positive self-reported perceptions of
learning. This experience report focuses on the design, classroom use, and
student perceptions of GenAI-PassLab rather than on controlled measurement of
learning outcomes, positioning the lab as a scalable proof of concept for
AI-enabled cybersecurity education.