Guidance
Plain and practical, for faculty.
Teaching & learning
Teaching is where AI's effects on the MSU community are often most visible and most consequential. Faculty discretion is the operating principle: you determine, and communicate, the role of AI in your courses.
Pedagogical discretion
A faculty member may prohibit AI use entirely, require AI use as a graded component, or specify any approach in between. The fact that the University provides AI tools does not substitute for course-level decisions and student accountability. Faculty retain the authority to restrict use.
Pedagogical decisions are honored across the full range of approaches. MSU supports faculty at every point on that range with pedagogical consultation, assessment design, and instructional resources through CITE and partner units.
Designing assessments
Assessment is the most-cited faculty concern. These are resources, not requirements or the only solutions; CITE maintains a growing catalog. Select any one to read more.
Process over product Assess drafts and revisions, not just the final file.
Assess the work behind the work (drafts, revisions, edit history, intermediate artifacts) rather than only final outputs.
AI-resilient by design Anchor assignments in local, specific context.
Design assignments that connect to specific course discussions, local context, or personal observation, where generic AI output is readily distinguishable from authentic student work.
AI as the object of study Grade the discernment, not only the deliverable.
Require students to generate multiple AI outputs, evaluate them critically, and submit their selection and reasoning. Assess the discernment, not only the deliverable.
In-class and handwritten assessment Still valid and valued; CITE can help with logistics.
MSU continues to support in-class and handwritten assessment as a valid and valued approach. Faculty may request institutional support for proctoring logistics through CITE.
On AI detection tools: MSU does not endorse reliance on AI detection software as the basis for academic integrity findings. Detection tools have known and significant false-positive rates, and their accuracy degrades as AI systems improve. If you suspect unpermitted use, rely on process evidence instead (drafts, version history, a conversation with the student) and route concerns through MSU's standard academic integrity process, not a detector score.
Your own use of AI in teaching
The same data rules that apply everywhere at MSU apply to teaching work. Student work and student information (papers, grades, advising notes, the details that go into a letter of recommendation) belong only in MSU tools you're signed in to with your NetID, and strip identifiers where you can. AI can help you draft feedback, but the judgment and the grade are yours: a person makes the final call on every grade.
Disclosure runs both ways. If you expect students to disclose their AI use, extend the same courtesy: tell students when course materials or feedback were produced with meaningful AI assistance.
Tiered AI use in courses
Communicate AI expectations clearly in your syllabus and course materials. CITE offers suggested syllabus language on its resources page. You are not required to use all three tiers: a course may operate entirely under one, or vary by assignment. Students follow the tier you set.
| Tier | What it means |
|---|---|
| Prohibited | No use of AI tools in any part of the assignment or course. Use is a violation of academic integrity expectations. |
| AI-Assisted (permitted with disclosure) | AI may be used to support the student's own work at stages of the process the instructor allows. Disclosure is required. Final work must reflect the student's own reasoning and accountability. |
| AI-Required (expected or graded) | AI use is expected or graded as a component of the assignment. The student demonstrates competence in working with AI, not avoidance of it. |
Baseline: mechanical assistance such as spell-check or autocorrect is allowed in any tier unless the instructor explicitly excludes it. Instructors who intend truly zero assistance (for example, handwritten in-class work) should say so in the assignment.
Research & scholarly activity
AI is increasingly part of research workflows: literature review, coding, analysis, drafting, editing. This is a brief advisory; it does not replace the requirements of sponsors, funding agencies, or MSU's research compliance offices, which control where they apply. Questions go to the Office of Research Compliance and Security.
Follow sponsor and agency requirements. Read each solicitation; NIH has specific rules.
Funding agencies are setting their own AI rules, and those rules vary. Principal investigators are responsible for reading each solicitation, award, and contract for AI-related terms before using AI in proposal preparation or funded work. NIH has issued specific guidance on AI use in applications (NOT-OD-25-132); investigators submitting to NIH should review it directly. Where a sponsor is silent, do not assume permission for uses that would disclose proprietary or unpublished content.
Protect restricted and controlled data. Export-controlled / ITAR data never goes in unapproved tools.
Export-controlled and ITAR-governed research data must not be entered into AI tools that have not been approved for that data. The consequences of mishandling controlled data are serious and fall on individuals. Any question about whether data is export-controlled or otherwise restricted should go to the Office of Research Compliance and Security before the data is used, not after.
Protect unpublished and proprietary work. Keep it out of personal AI accounts.
Unpublished manuscripts, data, code, and ideas entered into personal AI tools may be retained or used to train models. Keep unpublished and proprietary work in MSU-approved tools and out of personal AI accounts.
Disclose AI use and observe authorship norms. AI can't be an author; verify every citation.
AI tools cannot be authors; authorship requires accountability that only a person can hold. Disclose AI assistance in accordance with the norms of your venue (many journals now require a statement of AI use in the methods or acknowledgments) and verify any AI-generated text, citations, or references before submission. AI systems are known to fabricate citations.
Mind human subjects and IRB requirements. Stay within your approved protocol; FERPA / HIPAA apply.
Where research involves human subjects, AI use that touches identifiable participant data must be consistent with the approved IRB protocol and with FERPA and HIPAA where they apply.
The bottom line
AI is now embedded in the daily work of teaching, learning, research, and operations at Mississippi State. Faculty are designing both AI-forward and AI-resilient assessments. Students are using multiple AI tools. Research teams are weighing AI's role in scholarly work.
This is advisory guidance, not formal university policy. It does not constrain faculty pedagogical discretion or mandate AI use. It is written for all faculty, staff, and students, in every context where AI is used with university information and data, whether the tool is free, paid personally, or paid by the institution.
It is consistent with MSU's existing policies, including OP 01.10 (Information Security Policy) and OP 01.12 (Use of Information Technology Resources). Where this guidance and existing policy address the same matter, the existing policy controls.
Looking for something else? Students have their own version on the Students page, and general AI guidance for everyone at MSU is on the For everyone page. For MSU's full Information Security Program, see infosecurity.msstate.edu.
Last updated June 11, 2026 · Mississippi State University AI Hub