Reyal or Fa-ke!: Risks of Misidentifying AI-Generated Videos as Human-Made Among Computer Engineering Students at Bulacan State University

Abstract

This study investigates the ability of Computer Engineering students at Bulacan State University to distinguish between AI-generated and human-made videos. Using a quantitative descriptive correlational design, 105 third-year students participated in a video classification test and survey measuring accuracy, confidence, AI familiarity, and perceived risks. Results showed a moderate detection accuracy (mean = 3.27/5 or ~65%), indicating that even technically proficient students struggle to reliably identify synthetic media. While participants demonstrated high awareness of risks such as misinformation and deepfake misuse, this awareness did not significantly improve detection performance. The findings highlight a gap between AI knowledge and practical evaluation skills, emphasizing the need for integrating media literacy, ethical AI evaluation, and hands-on detection training in engineering education.

Ähnliche Arbeiten