Strict AI prohibitions can intensify students' moral conflict without preventing AI use¶

Claim · All claims
Evidence · 1 study · qualitative r3 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? A participant who wanted to follow a strict no-AI policy continued using AI and reported that the prohibition created internal moral conflict rather than helping her avoid AI. → Kim 2026

Evidence¶

Kim 2026¶

Kim, Jiyoon; Toyama, Kentaro; Kim, Sangmi; Carroll, John. (2026). "It's OK Because...": The Wild West of Student Rationalization of AI Use in Academic Writing. https://arxiv.org/abs/2605.29090 doi:10.48550/arxiv.2605.29090

q2 · i? · qualitative · r3

Interview case from the belief-versus-behavior analysis: participant P16 expressed a strong desire to obey a course policy prohibiting AI entirely, yet continued using AI. The article reports the prohibition "intensified their internal moral conflict", and P16 summarized that "all it's doing is creating moral conflict".

""I don't think that is actually helping me stay away from it, and it's actually creating conflict for me, because I'm breaking the rules. "(P16)"

Discussion¶

Related Claims¶

• Students' moral rationalizations of AI use are typically post hoc and often internally inconsistent — related
• By the 2025 admissions cycle, a majority of applicants to the studied public policy master's program submitted at least one primarily AI-generated essay despite an explicit prohibition — related
• Australian university policy environments yield mostly restrictive or indeterminate GenAI classifications, with no combination clearly permitted — related
• Institutional GenAI guidance is more pro-use than computing course syllabus policies, where 92% of syllabi set explicit permissions and half prohibit use outright — related
• Perceived institutional AI regulation differs significantly across STEM subjects, with mathematics and natural sciences perceived as prohibitive and computer science as permissive — related
• Universities converge on disclosure and attribution expectations as the condition distinguishing legitimate AI assistance from misrepresentation — related