This study examines the impact of institutional policies regulating the ethical use of generative artificial intelligence (AI) tools in programming education. Two distinct AI use policies were implemented in a five-week intervention with undergraduate engineering students: one emphasizing reflection through written self-assessment and another emphasizing transparency through the submission of full chat logs. Quantitative data from programming exams and survey responses, combined with qualitative analysis of open-ended feedback, revealed that both approaches led to comparable academic performance, suggesting that the reporting format itself has little effect on learning outcomes. However, differences emerged in students' attitudes and perceptions. The reflective policy promoted greater ethical awareness, honesty, and self-regulation, while the chatlog policy encouraged broader adoption of AI tools and stricter compliance with reporting requirements. These findings highlight the importance of clear institutional policies that balance flexibility, transparency, and ethical responsibility, positioning AI as a formative aid rather than a substitute for problem-solving. The study underscores the need to integrate ethical and digital literacy into programming curricula and to design educational frameworks that promote responsible and reflective AI use in higher education.
