The amount of AI-generated code in a submission was the dominant factor influencing students' ethical and rule-compliance judgments¶

Claim · All claims
Evidence · 1 study · associational r1 · q2
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? In a repeated-measures ANOVA over eight scenarios, all three coded factors (amount, understanding, stage) were significant (p<0.0001), but the amount of generated code in the final submission emerged as the dominant factor, with the largest cross-country differences in high/medium-amount scenarios. → Brian Harrington 2025

Evidence¶

Brian Harrington 2025¶

Brian Harrington, Irina Zlotnikova, Gayathri Nadarajan, and Samuel Ekundayo. (2025). Did Alice Do Wrong? Cross-Cultural Differences in Student Perceptions of Generative AI Use in University Computing Education. ACM Trans. Comput. Educ. 26, 1, Article 15. https://doi.org/10.1145/3776558

q2 · i? · associational · r1

Repeated-measures ANOVA with Scenario (eight levels) as within-subject factor and Country as between-subject factor. Amount of generated code had F=351.663, p<0.0001 for the unethicality prompt; Scenarios 4, 5, and 7 were exactly those coded High or Medium on amount, indicating students at both sites judged heavy AI-code submission more harshly.

"All factors are significant (p<0.0001), but the "Amount of generated code in final submission" emerged as the dominant factor influencing student perceptions, with Scenarios 4, 5, and 7 exhibiting the largest mean differences between students in CAN and KOR across both ethicality and rule compliance ratings."

Discussion¶

Related Claims¶

• Qualitative themes: efficiency and peer influence drive AI use, while confidence and rule ambiguity shape students' perceptions of dependence — related
• Canadian and Korean students did not differ significantly in overall opinions about GenAI's impact on academic integrity and learning quality, and academic progress did not affect scenario judgments — related
• Canadian students were more likely than South Korean students to judge GenAI-assisted coding scenarios as unethical — related
• Canadian students were more likely than South Korean students to judge GenAI-assisted coding as against institutional rules, despite functionally identical policies — related
• Students' perceptions of ethicality and of rule compliance were strongly and positively correlated, with rule ratings consistently higher than ethicality ratings — related
• Students' GenAI use is ethically negotiated, with decisions emerging from intersecting personal integrity, institutional rules, fear of punishment and practical constraints — related
• Students distinguish learning-support GenAI uses from substitutionary uses, with uncertainty highest for minimally modified AI content — related