Northern faculty are responsible for all course content, teaching decisions, and student assessment. Faculty are encouraged to use AI thoughtfully, communicate openly with students about its use, and ensure that AI-supported practices align with course goals, academic integrity, and university policies. Artificial intelligence (AI) offers new opportunities to support teaching and enhance student learning.
This page provides guidance, recommendations, and sample language to help faculty integrate AI effectively and responsibly into their teaching. The resources and examples may be adapted to fit the needs of individual courses and disciplines.
Frequently Asked Questions
What do I put in my syllabus about AI-generated content?
Northern recommends that all faculty adopt a clear, explicit AI policy that describes what is allowed, what is prohibited, how AI use should be documented, and how policy violations will be handled (see academic dishonesty and misconduct policy).
Importantly, AI expectations should be communicated early and often, not just stated in the syllabus. Discussing examples, assignment-specific expectations, and rationale throughout the semester helps reduce confusion and supports academic integrity.
Option 1: AI Prohibited
AI Use Policy: Prohibited
The learning objectives of this course require you to develop and demonstrate your own critical thinking, writing, analysis, and problem-solving skills. Therefore, the use of generative AI tools is not permitted for course assignments unless I explicitly authorize their use in writing.
Maintain evidence of your work, such as notes, drafts, outlines, and revision histories. Unauthorized use of AI-generated content may be treated as a violation of the university's academic dishonesty and misconduct policy and may result in academic penalties.
If you are unsure whether a particular use of AI is permitted, assume it is not and ask before using it.
Option 2: AI Allowed with Attribution
AI Use Policy: Allowed with Disclosure and Citation
Generative AI tools may be used in this course when they support your learning. You remain fully responsible for all content submitted for credit, including the accuracy of information, quality of analysis, and proper use of sources.
Any use of AI must be disclosed and cited. Include a brief statement describing: the AI tool used, how it was used, what portions of the work were AI-assisted, and what intellectual contributions remain your own.
Failure to disclose AI use when required may be considered an academic dishonesty and misconduct violation. AI-generated content should never replace your own judgment, analysis, or original thinking.
Sample disclosure statement: For this assignment, I used Microsoft Copilot to generate brainstorming ideas and suggest organizational improvements. All analysis, writing, source evaluation, and final content are my own.
Option 3: AI Permitted for Brainstorming and Revision Only
AI Use Policy: Limited Use
You may use AI tools for brainstorming topics and ideas, generating questions for further research, creating outlines, and grammar, proofreading, and style suggestions.
You may not use AI to draft substantial portions of assignments, write discussion posts, generate analyses, arguments, conclusions, or reflections that are submitted as your own work.
The purpose of this policy is to use AI as a support tool while preserving opportunities to develop your own thinking and communication skills. Any AI use must be acknowledged in an AI-use statement submitted with the assignment.
Option 4: Assignment-by-Assignment AI Use
AI Use Policy: Varies by Assignment
Because AI tools may support some learning objectives but undermine others, AI use will vary by assignment in this course. Each assignment will clearly indicate whether AI is not permitted, permitted with attribution, encouraged for specific tasks, or required as part of the assignment.
Students are responsible for following the instructions provided for each assignment and for documenting AI use when required. If assignment instructions and the syllabus appear to conflict, the assignment-specific guidance takes precedence.
Should I disclose my own use of AI tools?
Northern recommends that faculty members who incorporate AI tools into their teaching practice clearly communicate their use to students through their course syllabi and other appropriate course communications. Transparency regarding instructor use of AI promotes trust, clarifies expectations, and fosters an open learning environment. Here are two sample statements and disclosure language:
Option 1: To support student learning and enhance course design, the instructor may use AI tools for tasks such as developing instructional materials, generating discussion prompts, creating practice exercises, summarizing current research, and providing feedback suggestions. All course content is reviewed, evaluated, and edited by the instructor to ensure accuracy, relevance, and alignment with course learning objectives. The instructor remains responsible for all instructional decisions, grading, and assessment of student work.
Option 2: The instructor may use generative AI tools to assist in the preparation and delivery of course materials. Any AI-generated content is critically reviewed and revised by the instructor to ensure academic quality, disciplinary accuracy, and alignment with institutional and course standards. The instructor maintains full responsibility for all course content, pedagogical decisions, and evaluation of student performance. AI is used as a support tool rather than a substitute for professional expertise or instructor-student engagement.
Can I use AI detectors to check students' work?
AI detectors can be used as one source of information, but they should not be relied upon as evidence that a student used AI. Current AI detection tools are imperfect and can produce both false positives (human-written work flagged as AI-generated) and false negatives (AI-generated work not detected). Major vendors acknowledge that detection results are probabilistic and should be reviewed alongside other evidence rather than used as the sole basis for academic misconduct decisions.
See https://www.turnitin.com/blog/understanding-false-positives-within-our-ai-writing-detection-capabilities for more information.
Rather than relying on AI detectors, use assessments that make student learning and authorship more visible:
For example:
Multi-stage assignments: Require proposals, outlines, drafts, and revisions to document the writing process.
Oral defenses: Ask students to explain their arguments, evidence, and writing decisions.
In-class writing: Use short writing activities to establish a baseline sample of student work.
Reflection statements: Have students describe their process, revisions, challenges, and learning.
For more information, see “Other examples of scaffolding assignments for AI-resistant assignments”
How do I communicate expectations about acceptable AI use?
Start with Northern's academic dishonesty and misconduct policy and any applicable course or program guidelines. Students should understand that academic integrity standards apply regardless of whether work is completed with traditional tools or AI-assisted tools.
Academic Dishonesty and Misconduct Cheating and other forms of academic dishonesty and misconduct run contrary to the purposes of higher education and will not be tolerated. Academic dishonesty includes, but is not limited to, plagiarism, copy answers or work done by another student (either on an exam or an assignment), allowing another student to copy from you, and using unauthorized materials during an exam. Northern State Universities policies and procedures on academic dishonesty can be found in the NSU Student Handbook. The Board of Regents polices can be found in Board of Regents Policy 2:33 and Board of Regents Policy 3:4. The consequences for cheating and academic dishonesty are outlined in the above mentioned policies.
What are examples of assessments that incorporate AI intentionally?
The AI Assessment Scale (AIAS) is a five-level framework that helps educators decide what role AI should play in an assessment task and redesign the task so that decision holds up in practice. For more information, visit https://aiassessmentscale.com/
The following assignments shift the focus from having AI produce work for students to having students demonstrate critical thinking, evaluation, source verification, and disciplinary expertise while engaging with AI tools responsibly:
- Ask students to analyze AI-generated responses, identify inaccuracies, unsupported claims, biases, or missing perspectives, and explain how they would improve the output.
- Have students compare an AI-generated answer with peer-reviewed articles or course readings, evaluating differences in accuracy, depth, evidence, and credibility.
- Students can create quiz questions, case scenarios, or practice problems with AI and then assess the quality of the questions, revise them, and explain their changes.
- Students can prompt AI to produce opposing viewpoints on a topic, then critique the strength of those arguments and develop evidence-based responses.
- Students use AI to brainstorm, generate possible research topics, project ideas, or outlines, then justify which ideas they selected and why.
- Students document their prompts, evaluate the quality of the outputs, and reflect on how AI helped or hindered their thinking.
- Students verify AI-generated claims, citations, and references using credible academic sources and report any inaccuracies or fabricated information.
How can I design assessments that are less reliant on AI-detection and more "AI-resistant" by design?
The most resilient defense against AI misuse isn't a better detector, as detectors themselves can be unreliable, produce false positives, and are easily defeated. The more effective strategy for faculty to embrace is to design assessments where the process of doing the work is visible, personal, and challenging to outsource to AI. Below are four complementary strategies…rather than adopting all four strategies for each assignment, feel free to mix and match what might work best for the content delivery and assessments within the course.
Scaffold the Assignment: Proposal → Draft → Final: Breaking a single high-stakes assignment into staged checkpoints does two things: it creates a paper trail of a student's evolving thinking, and it makes it far harder to generate the whole thing in one AI session the night before it's due.
How to structure it:
- Proposal/topic pitch-Students submit a working thesis, research question, or project idea, along with why it interests them. Require a brief rationale connecting it to something discussed in class or to their own experience.
- Annotated outline or source list-Have students explain, in their own words, why each source matters and how it relates to their argument. This is hard to fake convincingly without doing the reading.
- Rough draft with reflection- Submitted before the final version, ideally with a short cover memo: "What's working, what are you unsure about, what did you change since the proposal?"
- Final draft -Graded not just on the finished product but on demonstrated continuity with the earlier stages.
Ask students to respond to specific feedback from the previous stage in the next submission, which can be difficult to simulate without having actually received and understood that feedback.
Varying the checkpoints from assignment-to-assignment so the pattern itself isn't gameable.
Authentic, Local, and Current-Events Prompts: Generic prompts ("Discuss the causes of the French Revolution") are exactly what AI systems handle best, because the internet is full of similar answers. Prompts that require specific, local, or very recent information are harder for an AI system to answer well without the student doing real work.
Approaches to try:
Localize it-Ask students to analyze a decision made by your institution, city council, or a dataset specific to your region, rather than a generic case study.
Use current events-Require engagement with something that happened in the last 1–4 weeks, such as a news event, a just-published study, a recent policy change. This also has the side effect of testing whether students actually did the AI-assisted research well.
Require primary interaction- Interviews, observations, lab data the student personally collected, or fieldwork notes anchor the assignment to something only the student has access to instead of all inputs the AI platform has been trained with over time.
Ask for personal application-"Apply this framework to a decision you personally made" or "connect this concept to your own major/career path" resists generic AI output because it requires details only the student has.
Build in constraints AI can't perceive- Reference a specific class discussion, an idiosyncratic rubric term, or an in-class example and require students to use it accurately, which provides a strong sense of whether they were actually present and engaged in the class.
In-Class Components: Nothing beats direct observation. Building graded work into class time doesn't eliminate AI reliance for out-of-class components, but it ensures at least part of the grade reflects unmediated student thinking.
Consider including the following in your courses:
In-class writing or problem-solving-short, timed responses (handwritten or on a locked-down device) that connect directly to a take-home assignment, so the two can be compared for consistency of voice, reasoning, and skill level.
Oral defenses / conferences-a 5–10 minute conversation where the student explains their argument, walks through their process, or answers follow-up questions about their own paper. This is one of the strongest signals available, since it's very difficult to fake genuine understanding under live questioning.
In-class peer review-students present drafts to peers and respond to questions in real time.
Low-stakes in-class quizzes on the reading/draft-quick (handwritten) checks that a student engaged with material before class, which also builds useful data on whether a submitted assignment matches the student's demonstrated understanding.
Portfolio and Process-Based Grading: Instead of grading a single artifact, grade the trajectory of work across a term. This rewards growth, makes any one AI-generated submission stand out as an outlier, and shifts the incentive structure away from "produce a perfect final product" toward "show your thinking and growth over time."
Portfolio elements to consider:
Working portfolios -students maintain a running collection of drafts, notes, false starts, and revisions, submitted together with the final piece. Grade partly on evidence of revision and iteration, not just polish.
Reflective/metacognitive components- include a short reflection with each submission: What was hard? What did you change and why? What would you do differently? Generic AI text struggles to answer these convincingly and specifically to the content within your course.
Process documentation-for research or writing-heavy courses, consider requiring version history (e.g., Google Docs revision history, track changes) or screen recordings of drafting sessions for select assignments.
Self-assessment against rubric-have students evaluate their own work against the rubric before submitting, and grade partly on the accuracy and thoughtfulness of that self-assessment.
End-of-term synthesis-a capstone reflection asking students to trace their own development across the portfolio, referencing specific earlier pieces. This is very hard to produce without having actually done the work throughout the semester.
What do I need to know about AI browsers, plugins, and AI-enabled wearables (e.g., smart glasses) during exams?
Establish clear expectations for technology use.
AI-enabled browsers, browser extensions, smart glasses, watches, and other connected devices create new considerations for academic integrity during proctored exams. Faculty should clearly communicate which technologies are permitted or restricted and consider what a technology is capable of doing, rather than focusing only on a particular device or product.
Consider accessibility and approved accommodations.
Some students rely on assistive technology, smart devices, digital resources, or connected devices for disability-related access or medical needs. Faculty should not independently deny the use of an approved assistive technology or accommodation because of concerns about its AI or connectivity capabilities. Instead, faculty should consult with Student Accessibility Services to determine an appropriate solution that provides access while maintaining the learning objectives and integrity of the assessment.
Work with Student Accessibility Services when technology creates testing concerns.
If an approved technology could provide access to outside information or create other test-security concerns, Student Accessibility Services can work with the student and faculty member to identify an alternative accommodation. Depending on the student's needs, this could include a reader, scribe, enlarged materials, approved digital resources, or another testing arrangement.
Be aware of limitations with lockdown browsers.
Lockdown browsers may create barriers for students who rely on assistive technologies or approved digital resources. Faculty should recognize that a standard lockdown environment may not work with every accommodation. Student Accessibility Services can help identify alternatives that maintain both accessibility and exam security. Technology Services may also be able to provide guidance regarding the capabilities or restrictions of specific browsers, extensions, and devices.
Include clear and inclusive proctoring language. Faculty may consider including language such as:
“The use of AI-enabled devices, smart wearables, browser extensions, or other technologies is not permitted during exams unless specifically approved by the instructor or authorized as a disability accommodation through Student Accessibility Services. Students who use technology for disability-related access should contact Student Accessibility Services and communicate with the instructor as appropriate regarding approved accommodations.”
What is agentic AI and why does it matter for teaching?
AI Definitions, Policies, and Contacts
For teaching, agentic AI represents a significant shift because it can complete substantial portions of a course independently (without human intervention), including researching information, organizing ideas, drafting content, revising work, and using other digital tools to accomplish complex tasks. As a result, instructors may need to rethink how they assess student learning.
Rather than asking whether students used AI, faculty should consider how AI was used and whether students are demonstrating the intended learning outcomes. Clear expectations, thoughtfully designed assignments, and assessments that require students to explain and apply their knowledge can help preserve academic integrity while preparing students to use AI responsibly.
Where can I get help building an AI policy or redesigning an assignment?
Link to # 1 (syllabus page/links to syllabus samples).
Center for Excellence in Teaching and Learning (CETL) at NSU
Center for Excellence in Teaching and Learning
Northern has a micro-credential for that! Check out AI in Teaching: Making Informed Decisions for Higher Education Faculty at https://d2l.sdbor.edu/d2l/le/discovery/view/course/2224866
This introductory micro-credential empowers faculty to make informed, intentional decisions about whether, when, and how AI aligns with their teaching philosophy, course learning outcomes, and program goals. Whether faculty choose to fully integrate AI, allow limited use, or restrict its use in certain learning activities, they will have a clear framework for making thoughtful, evidence-based decisions about AI in their teaching. Approximate time to complete is 3-5 hours.
Is it ethical to use AI in my course?
When determining the ethical nature of using AI in your course, it’s important to remember that instructors need to consider the purpose, context, risks, and consequences of a particular use of AI. AI should be evaluated as an educational tool, not simply a technological tool, that follows a human-centered approach in education. Preparing students for the world in which they will live and work is an ethical obligation instructors have. With that, instructors should consider the learning objectives they would like their students to achieve. By analyzing what students should learn or demonstrate, and then asking, ‘does AI support or interfere with that learning?’ instructors can gauge the degree with which they should integrate AI into their course(s).
What AI tool should I use? What am I allowed to use?
Copilot Chat
NSU provides a licensed version of Copilot Chat to staff, faculty, and students.
There are three ways to access Copilot Chat:
Go to MyNSU, then scroll down and click the link to Office 365. It will take you to https://www.office.com/. You may need to log in using your Northern account. On this site, you can use the web-based version of Copilot Chat.
Copilot Chat is available in Teams on the left-hand menu. If you don’t see it, open the three dots “…” first.
Copilot Chat is available as a standalone app installed on faculty and staff computers – “Microsoft 365 Copilot”.
O365 Copilot
Departments may purchase a licensed version of O365 Copilot for staff or faculty.
This provides additional features in Copilot Chat and allows the use of Copilot in many O365 apps, such as Outlook, Word, Powerpoint, and Excel.The cost is approximately $20 per month per license.
To purchase, submit a ticket and include the index code to be charged: https://td.northern.edu/TDClient/30/Portal/Requests/ServiceOffering/98/O365-Access-to-App. It may take up to two weeks for access to be granted.
What inputs are we allowed to put into AI?
The information faculty, staff, and students enter into an AI tool matters. Prompts, uploaded documents, images, datasets, and other information provided to an AI system may be processed or stored outside of university-controlled systems, depending on the tool and account being used. As a general guideline:
Use public or non-sensitive information.
Information that is already publicly available or does not contain confidential, protected, or personally identifiable information is generally the safest type of content to use with AI tools. Per BOR Policy 7.9, restricted, regulated, internal, or institutional data should never be used with unapproved AI tools.
Protect student information.
Do not enter personally identifiable student information, grades, student records, accommodations, advising information, or other education records into an AI tool unless the university has specifically approved the tool and its use for that type of information. FERPA and university privacy requirements still apply when using AI.
Protect confidential or sensitive university information.
Do not enter confidential personnel information, financial information, passwords, private institutional data, unpublished research, or other restricted information into public AI tools.
Remove identifying information when possible.
If AI would be useful for a task, consider whether the information can be anonymized first. For example, an instructor could ask AI to help develop feedback on a sample assignment without including the student's name, ID, or other identifying information.
Consider what you are uploading, not just what you are typing.
Uploading a document, spreadsheet, image, recording, or dataset is also providing information to the AI system. Review files for restricted, regulated, internal, or institutional information before uploading them.
Is AI the right thing to implement in my course?
It depends! Northern has a micro-credential designed to help you answer this question! The “AI in Teaching: Making Informed Decisions for Higher Education Faculty” micro-credential empowers faculty to make informed, intentional decisions about whether, when, and how AI aligns with their teaching philosophy, course learning outcomes, and program goals.
Participants will explore the capabilities and limitations of generative AI, examine its potential impact on teaching and learning, and evaluate both the opportunities and challenges AI presents in their disciplines. Faculty will assess the readiness of their courses and programs for AI integration, consider issues related to academic integrity, ethics, equity, and student learning, and explore strategies for designing AI-appropriate assignments and assessments. Whether faculty choose to fully integrate AI, allow limited use, or restrict its use in certain learning activities, they will leave with a clear framework for making thoughtful, evidence-informed decisions about AI in their teaching.
Is ok not to implement AI?
Yes! Not every course needs to incorporate AI tools, and choosing not to use them can be a well-reasoned pedagogical decision rather than a gap.
Some considerations that may justify this choice as an instructor:
- Foundational skill development- In many disciplines, students need to build core competencies through direct, unaided practice before AI tools are introduced. Skills like critical analysis, original argumentation, and problem-solving often develop more durably when students first struggle with them independently. Introducing AI too early can short-circuit that struggle, which is often where the actual learning happens.
- Discipline-specific reasoning- Some fields have assessment structures (e.g., in-class exams, oral defenses, studio critiques, clinical skills checks) where the point is to verify what the student themselves can do. Other fields may be in the early stages of establishing norms for AI use, and it's reasonable to wait until those norms, along with the tools’ reliability in that domain, are better established. Further, some fields’ ethical stance on AI may preclude an instructor from utilizing the tools.
- Course goals and instructional design- If the course's central learning objectives are about process (e.g., learning to write, learning to code from first principles, learning to think through a proof) rather than output, AI use can work against those objectives even when it improves the output.
- Academic integrity and assessment validity- Instructors are responsible for being able to fairly and accurately assess student learning. If an instructor hasn't yet developed a way to do that in an AI-permissive environment for their course, opting out until they have is a reasonable and defensible choice.
- Equity and access- Not all students have equal access to, comfort with, or fluency in AI tools. An instructor may reasonably choose to level the playing field by keeping the course AI-free until access and preparation are more consistent.
Can I require AI use in my course?
Yes, thanks to academic freedom, faculty often have latitude in designing assignments and selecting learning technologies, BUT requiring AI depends on institutional policy, accessibility requirements, and the learning objectives of the course.
If you want to require the use of AI, please consider:
- Student Learning: Does AI support or interfere with student learning? Academic freedom should not come at the cost of student learning.
- Accessibility and equity: Students must have reasonable access to the required tools.
- Privacy and data security: Some institutions restrict the use of certain AI tools because of student data privacy concerns.
- Transparency: Best practice is to explain why AI is being required, how it supports learning outcomes, and how students will be evaluated. Link to sample syllabus statements.
- Institutional policy: Be sure to follow SDBOR, university, college, or department policies regarding AI use.
- Link to Student Handbook – Student Academic Misconduct Policy and Student Grievance Procedures.