Source: arXiv

Paper

Did Alice Do Wrong? Cross-Cultural Differences in Student Perceptions of Generative AI Use in University Computing Education — Harrington, Zlotnikova, Nadarajan, Ekundayo (submitted 22 Jul 2026)
ID: arXiv:2607.19699

What it claims

Generative AI in computing classes is not only a plagiarism tool debate — it is a culture debate. This scenario-based survey (Fall 2024) compared how students at Canadian and South Korean universities judged the ethicality and rule compliance of AI-assisted coding practices.

Canadian students were consistently more likely to call GenAI use both unethical and against institutional policy than Korean students — even when the paper describes institutional policies as functionally identical. Differences held across nearly all scenarios (Mann-Whitney U and correlations reported). The factor that most strongly shaped ethical judgments was how much AI-generated code was incorporated into the assignment. The authors interpret results through Hofstede dimensions (power distance, individualism, uncertainty avoidance) and argue that equitable AI integration in education must be culturally responsive.

The breakdown

The wire takeaway is “students disagree about ChatGPT.” The useful split is finer. First, the comparison is Canada vs South Korea under policies the authors describe as functionally identical — so the gap is not “one campus forgot to write a rule.” Second, judgments move most with how much AI-generated code lands in the submission, which is exactly the gray zone where handbooks are vague and peer culture fills the silence. Third, the Hofstede framing is interpretive scaffolding, not a moral ranking of nations: it is a hypothesis about why the same scenario feels like honesty in one classroom and misconduct in another.

Why readers outside the lab should care

If you study in Korea, hire Korea-trained juniors, or run a dual-campus coding program, “our handbook already bans unauthorized AI” is not enough. The same printed rule can sit on two moral maps. Korean classrooms may normalize help that Canadian peers still label cheating — and expatriate faculty who import home-campus norms will misread what students think is fair.

For Korelay readers: treat university AI policies as behavior contracts that need local social proof, not universal ethics.

What travelers and expats should watch

• Do ask your Korean department for scenario examples (how much AI paste is OK) — not only a one-line ban.
• Do assume a Canada/US-trained sense of “obvious cheating” may be stricter than local peer norms; verify before grading or accusing.
• Don’t equate “policy text matches” with “students will judge cases the same.”
• Expect preprint limits: survey site details and full tables live in the PDF; this brief cites the abstract framing.

Context

Read this as evidence that AI integrity is culturally loaded, not as proof Korean students “cheat more.” Same rules, different moral temperature — design guidelines for the campus you are actually on. For Korelay’s overseas reader, the behavior update is narrow: before you accuse, grade, or import a home-campus AI ban into a Korean computing course, run the Alice-style scenarios with local students and staff. If their answers diverge from yours, the handbook was never the whole policy.

Source

arXiv:2607.19699 — abstract and framing cited; open the OA PDF for methods, scenarios, and statistics. Do not republish the PDF.