Shanghai Jiao Tong University

SJTU’s2025 trial framework classifies uses as prohibited, limited, encouraged or open; students must follow each course’s AI rules. Teachers publish scope and citation formats through syllabus/class/grading rules. Categories are not permission for every assignment or exam.

Students should mark and cite necessary information about AI tools; the instructor explains applicable citation templates and permitted uses. Citation does not authorize prohibited use.

The trial framework forbids privacy violations and information theft in AI development/use contexts and requires school approval or filing when using school public data. It does not establish a full vendor-training or retention matrix.

SJTU Xflops documents a locally deployed DeepSeek service supported by Xflops/Network Information Center, with named models. Eligibility, price, vendor/log retention and task authorization are not established by the listing.

The2025 graduate academic standards have a separate research/degree misconduct and review framework for current postgraduates and prior SJTU postgraduate degree holders. They do not grant blanket AI thesis/publication permission.

The2024 integrity code states five-working-day objections after notification and refers disciplinary appeals to the student appeal rules. Separately,2025 discipline and2017 appeal texts use different deadlines. Their cross-document priority remains unresolved here; preserve the decision and seek the applicable route rather than merging clocks.

The2024 academic-integrity code supersedes2021. Dishonesty in an assignment/lab assessment results in zero for that component; repeated dishonesty in one course gives a course zero. The retake maximum60 applies when dishonesty led to disciplinary punishment, not every AI use. Joint programmes may have separately filed rules.

The2025 disciplinary process gives students an opportunity to respond; a hearing may be requested within three working days from receipt of the proposed disciplinary decision. This is a proposed-decision hearing clock, not the final-decision appeal deadline.

Teachers carry primary responsibility for AI-supported design, explain use scope and risks, and provide testing/filing where necessary. This does not approve every external product.