Abstract

This study explores the potential of leveraging generative systems — capable of producing texts, dialogues, and pedagogical examples — to educate individuals on making more secure decisions when using digital services. The research is situated at the intersection of generative capability, adaptive pedagogy, and cyber awareness; while generating a new explanation represents a technical function, selecting an explanation tailored to the learner's needs and evaluating its impact on comprehension constitutes a pedagogical decision. Therefore, the deployment of such systems requires a clear learning objective, a trusted knowledge source, and governance rules that allow for decision review, prior to any discussion of improving awareness or behavior. The significance of this relationship becomes evident when a learner receives a digital message requesting urgent action, encounters a request to share data, or uses an automated assistant in a professional context. In such situations, knowing the definition of phishing or privacy is insufficient if the learner cannot apply it to the decision at hand. Adaptive educational dialogue may enable tracking the learner's interpretation, clarifying the rationale behind their choice, and subsequently providing another example that addresses the same error. This work examines the conditions for constructing such a practice as a proposed strategy that can be examined later, without equating linguistic persuasion with the validity of security guidance. The National Institute of Standards and Technology (NIST) guide for building cybersecurity and privacy learning programs provides a foundation for organizing learning according to needs and roles and evaluating the program throughout its lifecycle. It is used here to distinguish between participation in training and achievement of the learning objective, while noting that it is a guide issued in a U.S. institutional context and not a binding policy for the Police College in the United Arab Emirates. In the UAE context, the government's presentation of the Charter for the Development and Use of Artificial Intelligence provides principles related to privacy, human oversight, and transparency. Furthermore, the official statement published via the Emirates News Agency (WAM) on January 11, 2026, reflects the interest of the UAE Cybersecurity Council in raising awareness of risks associated with AI-enabled fraud. This study utilizes these two sources to understand the context of prevention and governance, and does not infer from them the existence of an adaptive program implemented at the Police College or an evaluation of its outcomes (UAE Government, n.d.; WAM, 2026).