Resources, Prompting Strategies, and Workflow Integration for Faculty.
Through your OSU credentials, you have access to Copilot Chat — a web-based AI assistant with enterprise data protection. This is the free tier included with educational licenses that integrates with Outlook, Word, etc.
What this means: Copilot Chat works like ChatGPT or Claude — you copy/paste content in, get responses out. The main advantage is that your data isn't used to train the model. For most tasks, the free tiers of any major AI tool will work equally well.
The prompting skills in this guide work across all major AI assistants. Faculty and Students should experiment with multiple tools to understand their strengths.
Best for: Institutional data protection
Available via OSU login. Use this when working with student information or institutional data. Enterprise data protection means your inputs aren't used for training.
Faculty: Prefer this tool when handling any student-related content.
Best for: General-purpose tasks, largest ecosystem
The most widely used AI assistant. Strong at creative writing, coding help, and general knowledge. Free tier is capable; paid tier adds image generation and more.
Most students already have exposure to this one.
Best for: Long documents, nuanced writing, reasoning
Handles very long documents well (200K+ tokens). Known for thoughtful, well-structured responses and strong performance on complex reasoning tasks.
Good choice for analyzing lengthy reports or papers.
Best for: Google Workspace integration, current info
Connects to Gmail, Drive, Docs if you use Google Workspace. Has real-time web access built in. Generous free tier with strong multimodal capabilities.
Natural fit if you're already in Google's ecosystem.
Best for: Research with citations
Think of it as "AI-powered Google." Searches the web in real-time and automatically cites sources. Excellent for fact-finding, literature reviews, and teaching students about verification—every answer shows where the information came from.
Particularly valuable for capstone research where sourcing matters.
Best for: Deep analysis of your own documents
Upload PDFs, articles, or YouTube videos and ask questions specifically about that content. Creates study guides, summaries, and even AI-generated "podcast" discussions of your materials. Answers are grounded only in what you uploaded—no hallucinations from general knowledge.
Great for students synthesizing multiple sources for capstone projects.
AI tools are powerful, but they have significant limitations. Understanding these is essential for responsible use.
AI confidently generates plausible-sounding but completely false information. This includes fake citations, invented statistics, non-existent people, and fabricated quotes.
Always verify:
Some tasks are poorly suited to AI assistance, even with verification.
Treat AI output as a first draft from an enthusiastic but unreliable intern, not as finished work.
Good workflow: AI generates → You verify → You edit → You own the final product. The time savings come from having a starting point, not from skipping review.
AI works best when you already know the subject well enough to catch errors.
Key insight: The more you know about a topic, the more useful AI becomes—because you can spot and correct its mistakes. Using AI for topics outside your expertise is high-risk.
Don't just ask a question; tell the AI who to be. This shapes tone, vocabulary, and level of detail.
"You are a community college advisor who is encouraging but realistic. Write an email to students explaining the consequences of missing the financial aid deadline."
If you want a specific output, show the AI examples of what "good" looks like. Give 2-3 input/output pairs.
"Write feedback comments for student assignments. Match this style:
Example 1: 'Your thesis is clear, but paragraphs 2-3 need stronger transitions. See my inline notes.'
Example 2: 'Strong analysis of the case study. Consider adding one more supporting example in section 4.'
Now write feedback for a student who had good ideas but poor organization."
Be explicit about how the data should be presented to save time on reformatting.
"Create 10 quiz questions about HVAC refrigerant handling. Format as a table with columns: Question | Correct Answer | Three Wrong Answers (pipe-separated) | Difficulty (Easy/Medium/Hard)"
For complex logic or calculations, ask the AI to show its work. This reduces errors and helps you verify the output.
"A patient's medication dosage needs adjustment. Think step-by-step: First, identify the current dosage. Second, calculate the weight-based adjustment. Third, check against maximum safe limits. Show your reasoning at each step."
The first result is a draft. Talk back to refine it. This iterative process often produces better results than trying to write the perfect initial prompt.
"That's too formal. Rewrite it for students in an automotive technology program—keep it professional but more conversational."
Practical applications with estimated time savings. Remember: always verify AI output before using.
Paste an assignment description and ask for a grading rubric with detailed descriptors for each performance level.
Example: "Create a rubric for evaluating welding bead quality with criteria for appearance, penetration, and safety compliance."
Ask the AI to generate three plausible but incorrect multiple-choice answers to go alongside your correct answer.
Example: "For a phlebotomy exam, create distractors for: 'The correct order of draw starts with blood culture bottles.'"
Get creative explanations that connect abstract concepts to students' existing knowledge and interests.
Example: "Explain network subnetting using an analogy to apartment building addresses for IT students."
Ask for a class session outline that balances instruction, hands-on practice, and assessment.
Example: "Create a 2-hour lab session outline for teaching IV insertion, including demo, practice stations, and competency check."
Practice challenging conversations (grade disputes, academic integrity, etc.) by asking the AI to role-play the student.
Example: "Act as a student who is upset about failing a clinical competency check. I'll practice explaining the decision."
Draft step-by-step procedures or checklists for technical processes, then verify against industry standards.
Example: "Create a pre-operation checklist for CNC machine setup. I'll verify against our shop's specific equipment."
External "Copy/Paste" strategies that do not require special plugins.
Create quizzes in bulk.
Faster grading feedback.
Professional pages without coding skills.
Clearer instructions.
Guidance for addressing AI use in your courses. Clear expectations prevent confusion and integrity issues.
Consider which approach fits your course goals:
Prohibited
"AI tools may not be used for any assignments in this course. All work must be entirely your own." Note: This approach is viturally impossible to enforce as commonplace tools increasingly embed AI features. Embedded AI is ubiquitous in most commonplace tools from Google Search to Microsoft Word.
Limited Use
"AI may be used for brainstorming and outlining only. Final written work must be your own. Disclose any AI use." Note: This, too, is difficulty to enforce. Well-intentioned students may not even know the extent to which less obvious AI systems supported their work
Permitted with Disclosure
"AI tools may be used as a writing assistant. You must disclose how AI was used and you remain responsible for accuracy." Note: This approach legitimizes AI use and helps develop student skills.
Frame AI as a tool that requires judgment, not a shortcut:
AI detection tools have significant false-positive rates and can flag human-written work. They're generally not reliable enough for academic integrity decisions. Focus on clear policies, authentic assessments, and open conversations instead.