Uncovering Hidden Perceptions: Ethical Awareness and AI Use Through Sentiment Analysis in Engineering Master’s Students

Abstract

Abstract This study investigates ethical awareness and academic integrity among master’s engineering students in the context of Artificial Intelligence (AI)-assited writing. An exploratory mixed-methods approach was combined with descriptive demographic data and sentiment analysis of student’s open-ended responses to a structured ethics questionnaire. The aim was to uncover nuanced emotional and cognitive patterns in student’s reflections on plagiarism, citation practice, AI use, and intercultural expectations regarding academic integrity. Sentiment analysis was conducted using the Orange Data Mining and the VADER model, enabling reproducible classification of responses into emotional polarities (positive, negative and neutral), supported by word clouds and sentiment heatmaps. Key findings include the influence of age, gender and academic stage on ethical perceptions: younger students adopt AI tools more frequently, while older students with industry experience tend to favor traditional research ethics. Female students demonstrate greater caution and stronger ethical concerns regarding AI use, whereas male students engage less in citation ethics discussions but report freer AI usage. Early-stage dissertation students often struggle with ethical referencing and methodology, while advanced students prioritize data integrity and informed consent. Students with prior ethics training show greater confidence in ethical decision-making compared to those relying more heavily on AI for citation management. Although most responses revealed neutral sentiment, the analysis revealed concerns about group-based misconduct, uneven institutional training, and ethical uncertainty surrounding AI use. Word clouds and sentiment heatmaps identified subtle affective patterns often overlooked in traditional methods. These insights guide the improvement of engineering programs, aiming to develop reflective, culturally responsive, and AI-aware ethics education.

Ähnliche Arbeiten