This paper presents the results from the second year of a research project investigating the integration of Generative Artificial Intelligence (GenAI) in online assessment within a Distance Education context. Following the institutional policy which promotes the thoughtful and transparent use of AI rather than its prohibition, this study explores how pedagogical strategies can foster academic integrity and literacy.

Building on the first year's methodology—where students submitted a GenAI Use Self-Report alongside their assignments—the second year introduced a strategic shift. While the first assignment maintained the AI use Self-Report model, the second assignment made the use of GenAI tools mandatory and integrated into the task, requiring students to provide a balance reflexion on the tools’ contributions to their work. This evolution allowed the researchers to address new questions regarding patterns of use, emerging themes in student reflections, and the comparative impact of these two integration strategies on assessment products.

Results indicate a significant increase in reported GenAI use, although the preference for specific tools remained relatively stable compared to initial findings. While student use cases showed some variation, a somewhat surprising finding occurred: the mandatory integration strategy did not result in enhanced quality of the final assignment products. This paper discusses these results in light of existing literature, proposing explanations for why deeper integration did not immediately translate into higher academic performance. The findings offer critical insights for institutions seeking to develop robust assessments that effectively incorporate GenAI while maintaining pedagogical validity and ethical integrity.