USC

Students must confirm with the instructor what AI use, if any, is permitted before using it. Instructors control course-specific rules.

编码说明CODING NOTEStudents must clear AI use with the instructor before using it; permission is a precondition, not a default.

52 校中 19 所与此相同 · 该条主流口径：默认禁止（19/45）19 of 52 take this position · most common: Prohibited by default (19/45)

Instructors control course-specific rules. USC’s provision of ChatGPT Edu does not change or override those rules.

编码说明CODING NOTECourse rules control; institutional tool provision does not loosen them. University data classifications are stated to remain separately in force.

52 校中 12 所与此相同 · 该条主流口径：教师可定但有底线（33/45）12 of 52 take this position · most common: Instructor rules within limits (33/45)

USC also directs users to be transparent about where and how AI contributed to written work. Attribution does not cure use that the instructor prohibited.

编码说明CODING NOTETransparency about where and how AI contributed is directed, but disclosure never legitimizes use the instructor banned.

52 校中 20 所与此相同 · 该条主流口径：强制披露（20/46）20 of 52 take this position · most common: Required (20/46)

Users must review output for accuracy, completeness, relevance, and appropriateness. USC highlights fabricated or misattributed quotations and citations as particular risks and expects human verification.

编码说明CODING NOTEMandatory review of output, with fabricated or misattributed quotations and citations singled out as risks.

52 校中 17 所与此相同 · 该条主流口径：须自行核实（17/38）17 of 52 take this position · most common: Verification required (17/38)

AI-authored work represented as the student’s own is plagiarism whether it is copied, nearly copied, or paraphrased.

编码说明CODING NOTEPassing off AI-authored text as one's own is named plagiarism, including near-copies and paraphrase.

52 校中 13 所与此相同 · 该条主流口径：学术不端（28/44）13 of 52 take this position · most common: Academic integrity violation (28/44)

Students should be informed before detection software is used, identifying information should be removed, and detector evidence alone is insufficient to determine responsibility.

编码说明CODING NOTEDetector output alone cannot establish responsibility; human review and context are called necessary, while notice to students before detection is advised.

52 校中 8 所与此相同 · 该条主流口径：不得作为唯一或主要证据（8/14）8 of 52 take this position · most common: Not sole or primary evidence (8/14)

Restricted confidential or confidential data—including personally identifiable information, health information, and protected education records—must not be shared with an AI tool without appropriate consultation.

编码说明CODING NOTENamed data categories are barred absent appropriate consultation; the bar covers prompts, uploads, and training data.

52 校中 20 所与此相同 · 该条主流口径：禁止上传受限数据（20/31）20 of 52 take this position · most common: Restricted data prohibited (20/31)

Sharing includes prompts, uploads, and training data. Users remain responsible for privacy, security, retention, and intellectual-property requirements.

编码说明CODING NOTEUsers are held responsible for intellectual-property requirements alongside privacy, security, and retention.

52 校中 25 所与此相同 · 该条主流口径：有明确规定（25/25）25 of 52 take this position · most common: Addressed (25/25)

USC’s institutional ChatGPT Edu environment protects USC data from use in OpenAI model training, but that protection does not displace university data classifications or course rules.

编码说明CODING NOTEUSC provides an institutional ChatGPT Edu environment; access to it does not authorize coursework use.

52 校中 19 所与此相同 · 该条主流口径：提供机构版工具（19/24）19 of 52 take this position · most common: Institution-provided tool (19/24)