Student Toolkit

Ryan Wright, Co-Director of AI Research at the University of Virginia, Special Advisor to the Provost for AI, and Rolls-Royce Commonwealth Eminent Professor of Commerce, has played an instrumental role in advancing AI literacy and engagement across the UVA community. Through his work on the AI @ UVA Exchange Substack, he has created a valuable platform for sharing insights, resources, and emerging developments in artificial intelligence, helping faculty, staff, and students navigate the rapidly evolving AI landscape.

AI will force us humans to double down on those talents and skills that only humans possess.

~ David Brooks, New York Times

Below is a list of contributors to the AI for Academic Excellence:

• Kiera Allison 
• Stefanie Conley 
• Roger Martin 
• Jingjing Li 
• Reza Mousavi 
• Chris Maurer 

Artificial Intelligence (AI) is one of the most powerful learning partners available to students today. Like any tool, its value depends on how you use it. AI can help you move faster, see new angles, and optimize your effort — the kinds of “shortcuts” humans have always embraced to work smarter, not harder. 

What matters most is whether those shortcuts serve your growth. Research is clear that ineffective AI use dulls critical thinking, weakens your ability to recall and combine ideas in original ways, fosters conformity, and risks snuffing out genius.  

By contrast, effective AI use sharpens your thinking, broadens your perspective, and amplifies your unique voice. 

The goal isn’t to avoid shortcuts. It’s to choose the right shortcuts — the ones that accelerate learning, not diminish it. 

Steps for AI Academic Excellence

Rule of thumb: Always check with your instructor. 

Before starting any academic work with AI, review the Generative AI Guidance for Faculty & Students (FAQs). This is the University’s official policy resource. It explains: 

• When and how you are allowed to use AI in coursework 
• Your responsibilities under the UVA Honor System 
• How faculty may set assignment-specific rules 
• Examples of acceptable vs. unacceptable AI use 

This document is important because it tells you the baseline expectations for academic integrity at UVA. The principles and practices in this Student Toolkit build on that guidance and translate it into practical steps for everyday use. 

For more Information, see Appendix A

Rule of thumb: Let AI be your assistant, not your author. If your work still sounds like you—your ideas, your reasoning, your voice—you’re using it well.

Before you begin using AI in your academic work, it’s important to ground yourself in a set of principles. These serve as your compass: they remind you that while AI can accelerate your process, the quality of your thinking, creativity, and integrity must remain at the center. 

The key is knowing when to trust, when to question, and when to verify. This comes best from experience.”  - Rewiring your Mind for AI, David Wood 

When you keep these principles in mind, AI becomes a powerful partner for growth rather than a shortcut that weakens your learning. Use them as a mental checklist each time you turn to AI for help. 

Lead with your voice — AI can assist, but your original thinking, creativity, and perspective must drive the work. 

Be transparent — Clearly disclose when and how you use AI. Misrepresenting AI-generated work as your own may be considered an Honor violation under UVA policy. 

Work iteratively — Use multiple prompts to brainstorm, refine, and stretch your ideas. Treat AI as a thought partner, not a one-click solution. 

Protect your privacy — Use only University-approved tools. Never upload your work or personal information to unverified platforms. 

Own the words — Rewrite AI output in your own style so your authentic voice comes through. 

Think critically — AI responses can be inaccurate, biased, or incomplete. Fact-check, cross-reference, and use sound judgment before incorporating them. 

Stay purpose-driven — Use AI in ways that deepen your learning, sharpen your skills, and advance your professional growth. 

When AI Doesn't Work

Sometimes AI responses will be incomplete, off-target, or biased. This doesn’t mean the tool is useless — it means you need to adapt. Employers like PwC note that adaptability, curiosity, and the ability to identify the strengths of different technologies are critical skills for future professionals. Here’s how to practice that mindset:  

Be iterative — Revise your prompt to add clarity, context, or constraints. Small changes often yield better results. 

Try a different angle — If one tool struggles, reframe your question (e.g., ask for an example instead of a definition). 

Switch tools — Different GenAI platforms have different strengths. Explore and compare outputs to see which aligns best with your task. 

Stay curious — Treat breakdowns not as failures but as opportunities to learn how the technology works—and where it doesn’t. 

Explore the Strength of Different GenAI Tools

Here are a few widely used tools and what they’re especially good at: 

Co-Pilot/ChatGPT — Strong generalist; useful for brainstorming, drafting, explaining concepts, and tutoring. 

Claude (Anthropic) — Great at reading longer documents, summarizing, and offering clear, structured reasoning. 

Perplexity — Best for fact-finding with citations; useful when you need sources and links. 

GitHub Copilot — Tailored for coding; generates, completes, and debugs code in multiple languages. 

Quillbot / Grammarly — Focused on writing clarity, grammar, paraphrasing, and style. 

Midjourney / DALL·E — Strong for creating visuals, images, and design concepts from text prompts. 

Just as you wouldn’t use Excel for graphic design, don’t expect one AI tool to excel at every task. Part of becoming “AI fluent” is learning which tool to use for which job. 

Rule of thumb: Match the tool to the task. Use AI where it adds value—brainstorming, testing ideas, or polishing—but let your own analysis, judgment, and creativity set the direction.

AI can support your work in many ways, but its strengths depend on the task. Use this guide to see where AI adds the most value — and remember, your critical thinking and originality must always lead. 

• Creative Assignments (Essays, Reports, Reflections)
  • Brainstorming fresh ideas, generating outlines, offering alternative perspectives, suggesting examples, and improving clarity and tone. 
• Case Studies (Critical Thinking & Decision-Making)
  • Stress-testing arguments, uncovering blind spots, generating counterarguments, refining structure, and exploring stakeholder impacts. 
• Coding / Technical Assignments
  • Drafting pseudocode, debugging errors, explaining code step-by-step, suggesting optimizations, and generating unit tests. 
• Presentations
  • Structuring slide decks, developing talking points, suggesting visuals or examples, and simplifying complex ideas for audiences. 
• Studying & Exam Prep
  • Creating practice questions, quizzing with flashcards, summarizing class notes, generating analogies, and explaining difficult concepts in simple terms. 
• Language Support
  • Checking grammar, adjusting tone, paraphrasing for clarity, translating drafts, and making writing more concise. 
• Collaboration & Project Management
  • Summarizing group discussions, dividing tasks, drafting agendas, writing status updates, and tracking deadlines. 
• Statistical Analysis & Data Analytics
  • Suggesting appropriate statistical tests, drafting example code (R, Python, Stata), explaining outputs in plain language, and proposing visualization options (always to be verified by you). 

Rule of thumb: Let the workflow guide your process, but keep your judgment and creativity in the lead. AI should support your learning, not replace it. 

Not all assignments are the same — writing an essay, solving a case study, coding an assignment, or preparing a presentation each calls for a different approach. The way you use AI should reflect those differences. 

See Appendix B for several example workflows

AI Across Your Learning Journey (Courtesy of Jill Mitchell) 

At McIntire, learning isn’t a straight line but a spiral progression — you revisit skills at deeper levels of sophistication. AI’s role changes with you: 

• Beginning Learner (Tutor): Use AI to explain concepts and provide practice. Example: clarifying the difference between assets and liabilities in accounting. 
• Developing Learner (Practice Partner): Use AI to test your understanding and give feedback. Example: quizzing on adjusting entries or helping prepare financial statements. 
• Advanced Learner (Collaborator): Work with AI to generate and refine output, moving from practice to production. Example: drafting a cost-volume-profit analysis or flexible budget scenario. 
• Expert (Co-Creator): Combine your expertise with AI to push boundaries and innovate. Example: analyzing a 10-K with AI to spot unusual ratios and design an audit strategy. 
• Skills at Risk of Automation: Recognize tasks AI can already handle (e.g., preparing basic financial statements) and focus your time on higher-order human skills — judgment, ethics, creativity, and strategy — that remain indispensable

Rule of thumb: The strong prompt gives the AI clarity and boundaries — which makes the output far more aligned with your needs.

Role • Task • Requirement 

This approach helps you get clearer, more useful responses by giving the AI context, a specific goal, and constraints. 

• Role — Tell the AI what perspective to take (e.g., professor, consultant, statistician). 
• Task — State what you want the AI to do (e.g., explain, compare, debug, summarize). 
• Requirement — Add specific conditions, limits, or desired style (e.g., “150 words,” “APA style,” “for a beginner audience,” “with 3 bullet points”). 

Example: Explaining Regression 

Tier 1 – Basic Context 

Prompt: “Explain regression.” 

• Useful when you only need a quick definition or refresher. 

Tier 2 – More Context 

• Prompt: “Explain regression as if I’m preparing for an MSBA analytics exam. Use a business example, keep it under 200 words, and give me a few practice problems. Don’t give answers until I’ve responded.” 

Effective when you need structured exam prep and targeted examples. 

Tier 3 – Even More Context 

• Prompt: “I am a student in the University of Virginia’s MSA program, and I learned about regression, but I’m struggling to apply it at work. Act as my mentor: first, ask me simple questions to check my understanding. Then, gradually increase the difficulty until I demonstrate mastery. Keep responses concise and include business examples. Give me feedback while I’m learning. Once I’ve demonstrated mastery, ask about my work context and help me explore how regression could apply.” 

Why Context Matters 

There’s no single “best” prompt length or style. It depends on: 

• What you need: definition, exam prep, tutoring, or job application. 
• Your situation: quick study before class vs. preparing for an internship project. 

See example prompts in Appendix C

Rule of thumb: If AI shaped any part of your process — ideas, structure, wording, coding, or analysis — acknowledge it briefly and specifically.

Always acknowledge your use of AI Tools: As the University’s Generative AI Guidance for Faculty & Students notes, you are responsible for clearly acknowledging any role AI played in your assignments. 

General Guideline (Adapted from MLA) 

You should: 

• Cite AI content whenever you paraphrase, quote, or incorporate text, images, data, or other material generated by it. 
• Acknowledge functional uses (e.g., editing, translation, brainstorming, statistical help, analytics) in a note, your text, or another suitable location. 
• Vet all sources and outputs since AI may misattribute, invent, or miscalculate. 

Practical Tips 

• Disclose all uses of AI, whether for ideas, structure, code, or data analysis. 
• Be specific about how you used it (e.g., “generated sample R code” vs. “ran regression analysis”). 
• Always double-check: AI’s math, stats, and analytics may be wrong, biased, or incomplete. Verify against reliable tools or references. 

Sample Disclosure Statements 

• “I used ChatGPT to brainstorm possible counterarguments in Section 2. I reviewed, revised, and integrated them into my own analysis.” 
• “I used GitHub Copilot to suggest alternative syntax for my code. All logic and debugging were my own.” 

See disclosure templates in Appendix D

Acknowledgements and Contributors

This Toolkit is the brainchild of Ryan Wright, Special Advisor to the Provost for AI, and Rolls-Royce Commonwealth Eminent Professor of Commerce. Through his work on the AI @ UVA Exchange Substack, he has created a valuable platform for sharing insights, resources, and emerging developments in artificial intelligence, helping faculty, staff, and students navigate the rapidly evolving AI landscape.

Other contributors include:

• Kiera Allison, Assistant Professor of Commerce
• Stephanie Conley, Director of Instructional Design, McIntire School of Commerce
• Roger Martin, Cohen Family Business Minor Bicentennial Associate Professor of Commerce
• Jingjing Li, Associate Professor of Commerce; Director of the M.S. in Business Analytics Program
• Reza Mousavi, Bank of America Associate Professor of Commerce
• Chris Maurer, Associate Professor of Commerce

Appendix A: Use of Generative AI Tools

The University provides official guidance for students on appropriate AI use in coursework. You must adhere to these expectations and any assignment-specific rules. When AI is permitted, use it as a learning aid rather than a substitute for your thinking. Summarize AI contributions in your own words, and include a brief disclosure (e.g., "I used [tool] for [task]; all analysis and final writing are my own."). Students are ultimately responsible for the work they submit under the UVA Honor System. See the Provost’s Generative AI FAQs 

Provost’s Generative AI FAQs 

Honor & AI 

Misrepresenting AI-generated work as your own or using AI outside the instructor’s stated guidelines may be considered an Honor violation. Students are responsible for understanding expectations and accurately representing their contributions. 

Honor Committee’s AI guidance and case process: Artificial Intelligence | Honor Committee 

Appendix B: Common Task Playbooks

Creative Assignments (Essays, Reports, Reflections) 

• WRITE FIRST: begin with your own outline, thesis, or rough draft. This ensures your voice and ideas lead the process. 
• USE: After drafting, use AI to brainstorm alternative perspectives, stress-test arguments, or generate counterpoints. 
• REPROMPT: AI for mashups, counterpoints, or missed angles. 
• SUMMARIZE & REVISE: Put AI’s suggestions into your own words and integrate selectively so the paper still “sounds like you.” 
• WRITE LAST: and revise, citing all real sources you use (never AI 'invented' ones). 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Case Studies (Critical Thinking & Decision-Making) 

• DRAFT: an outline of your recommendation and rationale. 
• MEET: with your team (if applicable) to align on initial ideas. 
• USE: AI to stress-test your logic, uncover blind spots, and generate counterarguments. 
• INTEGRATE: AI input selectively, rewriting in your own words. 
• VERIFY: every source or statistic AI suggests. 
• POLISH: style and grammar with Word tools or approved AI tools. 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Coding / Technical Assignments 

• DEFINE: & pseudocode first. Write the problem statement, inputs/outputs, acceptance tests, and pseudocode before using AI. 
• CHECK: with others. In a group, review the logic together before coding. 
• GENERATE: from pseudocode. Use AI only to draft from your pseudocode or stubs. Be prepared to explain every line. 
• TEST: & debug. Write unit tests (including edge cases), validate outputs, and remove hallucinated imports/functions. 
• ITERATE: with understanding. Accept AI refactors/optimizations only if you can justify the trade-off (e.g., performance vs. readability). 
• DOCUMENT: in your voice. Add comments/docstrings: purpose, assumptions, limitations. Tag AI-assisted blocks briefly. 
• PACKAGE: & disclose. Submit pseudocode, final code, tests, run instructions, and an AI-use disclosure. Respect licensing/IP rules. 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Presentations

• DRAFT: your own key points or storyline. 
• USE: AI to suggest slide structures, talking points, or visuals. 
• REPROMPT: AI for alternative ways to explain complex ideas. 
• SUMMARIZE: and refine in your own words to fit your delivery style. 
• PRACTICE: aloud — ensure the flow sounds natural and authentic. 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Studying & Exam Prep 

• START: with your own notes or readings. 
• USE: AI to quiz you with practice questions or to explain tough concepts in simpler language. 
• REPROMPT: AI to generate analogies, examples, or step-by-step solutions. 
• ALWAYS: cross-check with textbooks or lecture materials. 
• SUMMARIZE: AI explanations in your own study guide, using your own words. 

Language Support 

• DRAFT: in your own words, even if imperfect. 
• USE: AI for grammar checks, paraphrasing, or tone adjustments. 
• REPROMPT: for clarity improvements (e.g., more formal, more concise). 
• COMPARE: side-by-side and keep the version that reflects your voice. 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Collaboration & Project Management 

• DEFINE: roles, goals, and timelines with your group first. 
• USE: AI to generate project plans, checklists, or meeting agendas. 
• REPROMPT: AI for summaries of group notes or brainstorming sessions. 
• SHARE: outputs with your team for critical review (never accept AI results uncritically). 
• USE: AI to draft progress updates or status reports, then revise together. 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Statistical Analysis & Data Analytics 

• DEFINE: the problem. Write down your research question, hypotheses, and the type of data you’re working with. 
• PLAN: the approach. Use AI to brainstorm possible statistical tests, models, or visualization methods. 
• GENERATE: support code. Ask AI to draft R, Python, or Stata code — but run and validate it yourself. 
• CHECK: accuracy. AI may suggest the wrong test, drop variables, or misinterpret assumptions. Always cross-check against textbooks, class notes, or verified statistical guides. 
• INTERPRET: independently. Use AI to explain statistical outputs in plain language, then rewrite in your own words. 
• VISUALIZE: responsibly. AI can suggest chart types, but ensure they fit your data and tell the right story. 
• DISCLOSE AI USE: add the AI-use disclosure before submitting. 

Appendix C: Prompt Starters by Assignment Type

Use these prompt starters to get more effective responses from AI.  

Each follows the Role • Task • Requirement model: 

• Role = the perspective you want AI to take,  
• Task = what you want it to do,  
• Requirement = limits, conditions, or style. 

Creative Writing / Essays 

• Act as a writing coach (Role). Suggest 10 angles I could take on [TOPIC] (Task). Keep them concise, one line each, tailored for a general audience (Requirement). 
• Act as a professor (Role). Draft a one-page outline for an essay on [TOPIC] (Task). Include 3 main arguments and one example for each (Requirement). 
• Act as a debate opponent (Role). Provide 3 counterarguments to this thesis: [THESIS] (Task). Limit each counterargument to 2 sentences (Requirement). 
• Act as an editor (Role). Rewrite this paragraph (Task) to be clearer and more persuasive, in 150 words or less (Requirement). 

Case Studies (Critical Thinking) 

• Act as a risk analyst (Role). Identify 5 assumptions in my outline [OUTLINE] (Task). Rate each as low, medium, or high risk with a one-sentence justification (Requirement). 
• Act as a skeptical stakeholder (Role). Generate 3 objections to [RECOMMENDATION] (Task). Keep each objection under 40 words (Requirement). 
• Act as a strategy consultant (Role). Create a decision tree with 3 options for [PROBLEM] (Task). Show 2 trade-offs per option in bullet form (Requirement). 
• Act as an ethicist (Role). Suggest 2 ethical considerations for this case (Task). Keep responses under 3 sentences each (Requirement). 

Coding / Technical Work 

• Act as a software engineer (Role). Translate this pseudocode into Python (Task). Include type hints and inline comments (Requirement). 
• Act as a QA tester (Role). Write unit tests for this function (Task). Include at least 3 edge cases in pytest format (Requirement). 
• Act as a computer science tutor (Role). Explain the time and space complexity of this algorithm (Task). Keep the explanation under 150 words (Requirement). 
• Act as a senior developer (Role). Refactor this function (Task) to improve readability and maintainability, without changing behavior (Requirement). 

Presentations

• Act as a communication coach (Role). Suggest a 5-slide structure for a presentation on [TOPIC] (Task). Limit each slide to a title and 3 bullets (Requirement). 
• Act as a teacher (Role). Generate 3 analogies to explain [COMPLEX IDEA] (Task). Keep each analogy under 2 sentences (Requirement). 
• Act as a speechwriter (Role). Draft an opening statement for a presentation on [TOPIC] (Task). Limit to 75 words, engaging and clear (Requirement). 
• Act as a design consultant (Role). Recommend visuals for a presentation on [TOPIC] (Task). Provide 3 suggestions, each described in one sentence (Requirement). 

Studying / Exam Prep 

• Act as a tutor (Role). Create 5 flashcards based on these notes [NOTES] (Task). Provide the question on one side and the answer on the other (Requirement). 
• Act as a teacher (Role). Explain [CONCEPT] (Task) as if I’m new to the subject, using everyday examples (Requirement). 
• Act as an exam prep coach (Role). Generate 3 practice questions about [TOPIC] (Task). Include detailed answers under 100 words each (Requirement). 
• Act as a study partner (Role). Summarize this reading [TEXT] (Task). Limit to 5 bullet points under 15 words each (Requirement). 

Language Support 

• Act as a grammar editor (Role). Review this paragraph (Task). Correct grammar and maintain a formal academic tone (Requirement). 
• Act as a paraphrasing assistant (Role). Rewrite this passage (Task). Make it more concise while keeping the meaning intact (Requirement). 
• Act as a translator (Role). Translate this text into English (Task). Ensure it reads naturally in academic style (Requirement). 
• Act as a style coach (Role). Suggest 3 alternative ways to phrase this sentence (Task). Each under 15 words (Requirement). 

Collaboration & Project Management 

• Act as a project manager (Role). Summarize this meeting transcript (Task). Provide 5 key action items in bullet points (Requirement). 
• Act as a planner (Role). Draft a project plan for [PROJECT] (Task). Include tasks, roles, and deadlines in a table (Requirement). 
• Act as a team lead (Role). Write a status update email based on [NOTES] (Task). Keep it under 150 words (Requirement). 
• Act as an organizer (Role). Generate a preparation checklist for a group presentation (Task). Provide 8 items maximum (Requirement). 

Statistical Analysis & Data Analytics 

• Act as a statistician (Role). Suggest 3 statistical tests for this dataset [DATA] (Task). Explain why each fits in 2 sentences (Requirement). 
• Act as a data scientist (Role). Write Python code to run a regression on [VARIABLES] (Task). Comment each line (Requirement). 
• Act as a teacher (Role). Interpret these outputs [OUTPUT] (Task). Write the explanation in plain language, under 150 words (Requirement). 
• Act as a visualization expert (Role). Recommend a chart type for this dataset [DATA] (Task). Provide 2 labeling tips (Requirement). 

Other Resources: 

• Ethan Mollick’s Prompt Library  
• MIT Sloan’s Prompt Engineering for Innovation 
• AI Prompt Database 

Appendix D: AI Disclosure Quick-Reference Menu

When you use AI, always acknowledge it. Below are plug-and-play templates you can adapt depending on the type of assignment. 

Creative Writing / Essays 

• I used [AI tool] to brainstorm alternative perspectives. All arguments and final writing are my own. 
• I used [AI tool] to check grammar and improve clarity. All ideas and analysis are my own. 

Case Studies / Critical Thinking 

• AI assisted with [task, e.g., generating counterarguments] in section [X]. I reviewed, revised, and integrated the content in my own words. 
• I used [AI tool] to stress-test my recommendations. All conclusions and final reasoning are my own. 

Coding / Technical Work 

• I wrote pseudocode first and used [AI tool] to suggest improvements. I debugged, commented, and verified all code myself. 
• I used [AI tool] to identify potential bugs. All fixes, explanations, and final code are my own. 

Presentations 

• I used [AI tool] to outline potential slide structures. I refined the content and chose visuals myself. 
• AI suggested alternative talking points for [topic]. I revised and practiced them in my own delivery style. 

Studying / Exam Prep 

• I used [AI tool] to generate sample flashcards. I checked the accuracy and rewrote the answers in my own words. 
• I asked [AI tool] to explain a difficult concept in simpler terms. I compared it with my class notes to ensure accuracy. 

Language Support 

• I used [AI tool] to check grammar and phrasing. All ideas and arguments are my own. 
• I used [AI tool] to translate draft text into English. I reviewed and revised the translation for accuracy and tone. 

Collaboration & Project Management 

• I used [AI tool] to draft an initial project plan. My team and I adjusted roles, tasks, and deadlines ourselves. 
• AI helped summarize our meeting notes into action items. The team reviewed and confirmed the list together. 

Statistical Analysis & Data Analytics 

• I used [AI tool] to suggest possible regression models. I validated the code in R and interpreted the results myself. 
• AI generated Python code to produce a chart. I verified the accuracy and ensured it fit my dataset and research question.