When learning to code, the digitization of source code profoundly impacts students. While the Internet provides abundant resources that could facilitate learning, it also increases the risk of reducing learning to a mere copy-and-paste exercise, leading to minimal actual learning. Applying systems thinking, the article analyses a series of student conversations to understand the kinds of mental models used by students to engage with their studies. In identifying such mental models enables the selection of strategies to address deficiencies. The objective is to create an environment that prioritizes the understanding of underlying issues that lead to dishonesty. The aim is to implement early interventions and establish a social student-lecturer contract; to create a safe space, conducive for students to learn and develop programming skills. The goal is to establish an educational setting where the focus is not on dishonesty but on addressing the factors that lead to it. Results indicate that unhealthy mental models pose a significant barrier to coping with the academic demands of a university environment.