Generative AI is now part of how students write, research, code, and study — whether or not we plan for it. This guide is a practical, research-grounded starting point for any educator who wants to use AI thoughtfully, protect the integrity of learning, and prepare students for a world in which AI fluency is a baseline professional skill.
Work smarter. Learn smarter. Change the world for good.
Last updated: October 2026. This is a living document, written for the 2026 landscape and updated periodically. Tools and policies change quickly — treat specifics as a snapshot, not the final word. Its companion for students is the AI Student Guide, and for staff, the AI Guide for Classified & Counselors.
This guide is shared by Professor Gary Graves as a colleague-to-colleague resource. It is not official college or district policy, and it doesn't tell you what to decide: how you handle AI in your classes is your call. Whatever you choose, whether you welcome AI, limit it, or don't allow it, put your expectations in your syllabus and on each assignment. Students do their best work when they know where the lines are.
The goal here is not to police AI or to wave it through, but to teach with it deliberately. The instructors who navigate this best tend to do three things: they learn the technology well enough to set informed expectations, they redesign assessment around what they actually want students to be able to do, and they talk openly with students about when and how AI use is appropriate.
A handful of durable principles hold up even as the tools churn:
You don't need to be an engineer, but spend real time with current AI tools before you set rules about them. Your credibility — and your policies — depend on understanding what these systems can and can't do.
"No AI" and "AI encouraged" are both valid — what's not valid is silence. State the rule for each assignment in your syllabus and in the assignment itself, because the right answer differs from a reflective essay to a coding project.
Students will use AI in their careers. Showing them how to prompt well, verify outputs, and use AI critically is part of preparing them — and it reframes the conversation from cheating to competence.
If you use AI to draft a rubric, generate examples, or build a study guide, say so. Disclosing your own use normalizes honesty and signals that students can talk with you about theirs.
Shift points away from easily automated tasks (recall quizzes, formulaic essays) toward work that shows a student's own thinking, process, and voice. This protects learning better than any detector.
Generative AI still fabricates facts, citations, and quotes. Teach students that they own everything they submit — and that "the AI said so" is never a defense for an error.
Don't require students to enter personal or sensitive data into AI tools, and don't put their data into consumer tools yourself. Don't assume everyone has paid access: favor tools students can use free, ideally through their @fullcoll.edu login, which adds education-tier data protections.
AI can help draft feedback, rubrics, and course materials, but grades, academic-integrity findings, and accommodation decisions stay with you. Review anything AI produces before a student sees it.
AI has real costs in energy, water, and data. Use it where it adds learning value and model that judgment for students. Sometimes the right answer to “should I use AI for this?” is no.
Review the output: is it accurate, appropriate, inclusive, and right for its audience? Revise it so it reflects your purpose, judgment, and voice. Verify facts, quotations, citations, numbers, links, and policy details against reliable sources. The AI Student Guide uses the same three words, so students hear one message.
It helps to know what we're actually trying to build in students. Recent scholarship gives us a useful map. Polomoshnov, Masso, and Lobanova (2025) review digital, data, and AI literacy and find that all three share three interrelated dimensions:
Knowing how to operate the tools — and having a working mental model of how AI systems generate their outputs.
Evaluating AI outputs for accuracy, bias, and limitations, and questioning when a tool should be used at all.
The often-overlooked dimension: how we interact with, reason alongside, and make sense of AI as a thinking partner.
Their key insight for teaching: the critical and communicative-cognitive dimensions — not just button-pushing — are what turn AI use into genuine literacy. Assignments that only build technical skill miss most of the point.
For a more concrete checklist, Annapureddy, Fornaroli, and Gatica-Perez (2025) propose twelve generative-AI competencies that progress from foundational to advanced — spanning understanding how generative models work, prompt engineering, evaluating and verifying outputs, recognizing bias and limitations, and reasoning through the ethical and legal implications of use. You can treat these as learning outcomes: pick the few that fit your course and design toward them.
Pick two or three literacy outcomes you genuinely care about in your discipline — say, "verify AI-generated claims against primary sources" and "disclose and reflect on AI use" — and build them explicitly into an assignment and its rubric. Literacy grows when it's assessed, not just mentioned.
A ready-to-use module you can add to any course to build these literacies directly with your students. Faculty can review and test the materials directly from the website, or import the module through Canvas Commons and customize it locally for your courses.
Topics included. The module introduces students to:
The goal is not simply to teach students how to use AI tools, but to help them become informed, critical, and responsible users of AI technologies.
The most reliable response to AI is not detection — it's assessment design. Work that is personal, process-rich, and tied to the real world is both harder to outsource and more worth doing. Some approaches that travel well across disciplines:
Take one existing assignment and add a required reflection: "If you used AI, show your prompts, explain what you kept or rejected, and describe what you learned." This single change converts AI use from a hidden risk into visible, gradable thinking.
Detection tools are unreliable and can falsely accuse students — so lead with clear expectations and humane conversations rather than enforcement theater. When you suspect inappropriate use:
A shared scale removes guesswork. Students see the same four levels in the AI Student Guide, so “AI Level 1” on an assignment means the same thing in every course that uses it.
| Level | What students may do |
|---|---|
| Level 0: No AI | All work is done without AI tools. |
| Level 1: AI to learn and plan | AI may explain concepts, quiz the student, or help brainstorm, but no AI-generated text, code, or images appear in the submission. |
| Level 2: AI-assisted, disclosed | AI may give feedback, help revise, or help draft parts of the work. Students disclose how they used it and remain responsible for the result. |
| Level 3: AI as collaborator | Full use is expected. Students document their prompts, critique the outputs, and reflect on what they kept and changed. |
Using an AI agent or AI browser to complete coursework counts as AI use under whatever level the assignment sets.
Clear syllabus language prevents most problems. Adapt one of these to your course:
AI not permitted:
AI permitted with disclosure:
How students should disclose:
Modeling transparent use (yours):
Using AI responsibly means naming its real risks for students:
The most capable tools often sit behind paid tiers. Requiring them can disadvantage students who can't pay. Favor free, no-account options for required work — and remember AI can also level the field by acting as a tutor, translator, or accessibility aid.
AI models learn from human data and can reproduce biased or offensive content. Outputs are not neutral or authoritative; teach students to read them critically.
Assume anything entered into a consumer AI tool may be retained or used for training. Never require students to submit personal, identifying, or sensitive information to these systems, and hold yourself to the same rule: don't paste student submissions, names, grades, or rosters into consumer AI tools to generate feedback or grades. Student education records are protected by FERPA, and removing a name isn't always enough when other details could still identify the student. The same care applies to health information, accommodation details, passwords, and confidential college records. If you use AI with student work, use a district-supported tool signed in with your @fullcoll.edu account, and check with district IT or Distance Education when you're unsure.
AI can speed up feedback drafts and rubric alignment, but a person makes the final call on grades, academic-integrity findings, and accommodations. Read and edit AI-drafted feedback before it reaches a student, and tell students when AI helped you produce it.
Copyright questions around AI training data and outputs remain unsettled and continue to move through the courts. AI is generally not recognized as an author; current guidance is to disclose and cite AI as a collaborator following your discipline's standard (APA, MLA, etc.). Don't upload publisher materials, textbook content, or other copyrighted work into consumer AI tools unless the license allows it, and don't assume AI output is free of copyright issues.
Two newer issues belong in your course planning:
AI agents and AI browsers now take actions, not just answer questions: they log in to sites, open files, fill in forms, and click through pages. An agent can complete a Canvas quiz or post to a discussion on a student's behalf. State in your syllabus that using an agent to complete coursework counts as AI use under each assignment's level. For your own work, don't give an agent access to Canvas, email, or files that contain student records unless the district has approved that tool for that use.
AI isn't weightless. Data centers draw significant electricity and water, and image, video, and large reasoning models use far more computing power than a short text request. The question usually isn't whether one prompt is harmful, but whether a workflow is frequent, scaled, or heavier than it needs to be. Use AI where it adds value, prefer text-based and source-grounded tools over heavy media generation when they'll do the job, and invite students into the same judgment.
The tool landscape changes monthly. Rather than maintain a list that ages, here are durable categories to explore — and a deeper, maintained directory.
Sign in with your college account. Microsoft Copilot and Gemini for Education used through your @fullcoll.edu login keep your work under the college's data protections and out of public model training. Personal accounts don't, so use them only for non-college work.
Free tools are a great way to experiment and get curious. For teaching, though, the difference between free and licensed tools is real:
So ask. If a licensed tool would make a real difference in your teaching, tell your dean, Distance Education, or the District AI Workgroup what you need and why. Specific requests from faculty are how these tools get funded.
The companion to this page: which tools to use, copy-and-paste prompts, the four AI levels, and disclosure templates students can paste into their work. Link it from your syllabus.
The Chancellor's Office hub for systemwide AI guidance, initiatives, and professional development. Prefer system-supported and district-supported tools where they fit, because they tend to come with stronger data-privacy and accessibility protections than consumer tools. Before adopting a new AI tool for a whole course or program, check with Distance Education or district IT.
My curated, regularly updated directory of AI tools and resources for educators and students, organized by use case. Start here for current, vetted recommendations.
For students: AI courses, certificates, and the AI in Business degree — plus an AI assistant that answers questions about the programs.
Policy and citation guidance
Questions, or want to compare notes on teaching with AI? Email me at ggraves@fullcoll.edu or stop by during office hours. This guide grows through conversation — I welcome suggestions and additions.
Work smarter. Learn smarter. Change the world for good.