<![CDATA[In the classroom, rules regarding writing authenticity often sound simple—do not copy, cite sources—yet in practice, they are rarely straightforward. This case study captures how assignment originality screening was conducted in a single class at UIN Sumatera Utara ($n=33$) over a three-week period, during which the lecturer combined detection tools (similarity matching/AI) with manual verification and structured feedback. We collected policy documents, examination results from two manuscript versions (pre- and post-revision), responses to an 11-item Likert questionnaire, open-ended comments, and conducted brief interviews with the lecturer and several students. Quantitative analysis presents per-item summaries, 95% confidence intervals, thematic subscales, and composite score distributions; qualitative analysis synthesizes recurring themes from comments and interviews, specifically focusing on rule clarity, perceived fairness, and identifying what genuinely assists students when revising their writing. The results indicate a neutral-to-positive attitude toward the process (composite mean $\approx 3.47$), peaking at the item regarding "increased intention to write authentically" after receiving feedback, and scoring lowest on "trust in the lecturer’s final decision." Based on these findings, we propose more transparent methods for explaining the basis of the lecturer's decisions (e.g., displaying text snippets and their original sources), feedback formats that directly address "what went wrong and how to correct it," and simple guidelines for interpreting similarity reports/AI indicators. While we acknowledge the study's limitations—a single class and a three-week duration—the findings are sufficiently rich to suggest practical measures that can be immediately adopted in similar classroom settings.]]>
