Teach: AI in Practice

Sample Activities

• Flipped Evaluations (Basics Focus) use an AI tool to generate text passages or arguments relevant to your course content. Have students work in small groups to evaluate the AI's output, assessing factual accuracy, finding logical flaws, checking citation validity, and identifying what improvements must be made.

• Comparative Outlining (Critical Thinking Focus) asks students to draft a research paper outline independently using only their minds and course materials. Then, ask them to generate an alternative version using AI and compare the two. This exercise develops critical thinking as students discuss why their personal structure may make more sense than the statistical average generated by the machine.

• Prompt Iteration Logs challenge students to write and iteratively refine prompts to solve a complex coding error or math problem. Students submit their "Prompt Log," showing how they successfully navigated the tool and directed it to achieve the correct result.

Discipline-Specific Examples 

Health Sciences & Healthcare: In healthcare, AI literacy should focus on bias verification.

• Have health science students analyze clinical datasets or simulations generated by AI to identify embedded racial, gender, or geographic biases explicitly.

• Teach students to question diagnostic assumptions, protecting the human element of clinical care.

English & Humanities: In writing-focused courses, prioritize voice preservation.

• Discuss how generative AI relies on linguistic statistical averages, which can homogenize and standardize unique student writing voices.

• Teach students to use AI selectively as a structural editor rather than an author, preserving their personal expression and cultural identity.

Business & Marketing: In business courses, prioritize critical curation and market validation.

• Have business students use AI to rapidly generate market research reports, competitive analyses, or startup business plans based on specific case studies.

• Teach students to act as executive decision-makers by auditing the AI's output against real-world economic indicators, verifying financial projections, and modifying the strategies to account for unpredictable human market behaviors.

Overview

Asking students to create something new forces them to be active creators and directors of their own work. This shift often fosters intellectual curiosity and transforms AI into a collaborative tool for creative expression rather than a shortcut for completion.

Examples

1. Graphic Design / Marketing: "The Creative Director’s Pitch"

• • The Project: Students design a comprehensive branding campaign for a local non-profit.

  • The AI Role: Students use AI image generators to create rapid, iterative mood boards and concept art to pitch to their "client."

  • The Artistic Shift: Instead of grading raw images, faculty assess the student's artistic direction—how well they curated, refined, and blended AI elements into a cohesive, human-polished final campaign layout.

2. History / Literature: "Visualizing the Unseen"

• • The Project: Students analyze a historical event or a complex literary text and generate a historical reimagining or scene illustration based strictly on textual evidence.

  • The AI Role: Students input precise historical or descriptive details into a prompt to generate an accurate visual representation.

  • The Artistic Shift: The assignment evaluates the student's critical skepticism. Students write a companion essay detailing where the AI image succeeded and where it failed (e.g., historical anachronisms, cultural biases), demonstrating mastery of the source text.

3. Biology / Environmental Science: "The Speculative Evolution Gallery"

• • The Project: Students apply principles of adaptation and natural selection to design a hypothetical organism engineered to survive a specific climate change scenario.

  • The AI Role: Students translate biological constraints into text prompts to render their creature.

  • The Artistic Shift: Students act as scientific illustrators. The project relies on iterative refinement—students must constantly tweak their scientific descriptions and prompts until the AI outputs an anatomically plausible organism that aligns with real ecological principles.

4. Philosophy / Ethics: "The Bias Lens Exhibit"

• • The Project: Students explore abstract philosophical concepts (like justice, beauty, or power) or professional roles (like "CEO" or "doctor") through visual media.

  • The AI Role: Students intentionally prompt the AI using neutral terms, then analyze the systemic biases, stereotypes, and cultural assumptions embedded in the generated results.

  • The Artistic Shift: Students act as critical curators. They use the AI tool as a mirror for societal patterns, assembling a digital gallery that critiques the technology's underlying training data rather than accepting the images at face value.

Supported Tools & Guided Workflows

To execute these projects, students can leverage our supported institutional tools or exercise their own creative license by selecting alternative software that best fits their artistic vision.

Option A: The Image Generation Suite (Good for Projects 1–4)

• Gemini (Co-Creator & Illustrator): Students use Gemini to generate their initial project images. To push them past basic prompts, have them use Gemini iteratively—first to help them draft a highly descriptive, vocabulary-rich prompt, and then to refine the style, lighting, or accuracy of the generated image.

• Alternative Tools (Adobe Firefly, Canva, Midjourney): Students may choose outside tools to achieve specific artistic aesthetics. This tool choice can be turned into a learning point: have students justify why their chosen platform was better suited for their artistic vision than the institutional standard.

Option B: The Infographic Pathway (Best for Visualizing Data & Research)

• NotebookLM (The Structural Prototype): Instead of just researching, students can automatically transform dense source materials, research papers, or historical archives into a structured visual overview. In the NotebookLM "Studio" panel, they select "Infographic" to instantly generate a visual hierarchy of their data.

• Gemini (The Design Critic): AI-generated layouts can occasionally suffer from awkward spacing or jumbled text. Students can download their NotebookLM infographic as a PNG, upload it into Gemini, and prompt it as a design mentor: "Analyze the visual hierarchy of this infographic. How can I improve the spacing, color theory, or clarity to better communicate this data?"

• Canva or Adobe Express (The Final Polish): Equipped with NotebookLM's structured data and Gemini's design critique, students step into free design platforms to manually piece together, polish, and claim full creative ownership over the final, human-executed asset.

💡 TIP: Think of these AI tools like different artistic mediums (watercolor, charcoal, or digital paint). No matter the tool, the pedagogical requirements remain identical: they MUST engage in rigorous research, analyze their output critically, and transparently disclose their process.

Resources

How to Use AI for Media Creation provides an overview on how to create AI-generated media, including images, audio, and video for multimedia projects (University of Alabama, 2024).

AI Image Generation: Techniques and Tools outlines tools and approaches instructors can use to create custom visuals that engage students and enhance learning (University of Alabama, 2024).

AI for Image Generation in Course Design explains how to use AI for image generation and how AI-generated images offer new opportunities to engage students in critical thinking and creative projects across disciplines (University of Alabama, 2024).

Overview

AI can assist in research by automating tasks, facilitating knowledge discovery, and enhancing efficiency. You can use AI tools for brainstorming, literature searching, data analysis, and writing assistance. AI can help with tasks like generating research questions, finding relevant articles, and summarizing complex information. 

How to effectively use AI for research

1. Define Your Research Goals: Clearly outline your research objectives and questions. This will guide your AI tool selection and usage. 

2. Choose the Right AI Tools: Select AI tools that align with your specific research needs. Choose tools that provide links to literature.

3. Leverage AI for brainstorming, refining the research question, locating relevant sources, data analysis, and text editing/refining.

4. Verify, Verify, Verify: Be mindful of data quality and potential biases in AI-generated results. Always validate AI output with your knowledge and expertise. 

5. Maintain Human Oversight: Don't rely solely on AI. Human oversight is crucial for ensuring accuracy, integrity, and validity. 

6. Consider Ethical Implications: Be aware of ethical considerations, such as plagiarism, data privacy, and citing AI-generated content. 

7. Be Mindful of Limitations: AI tools should ONLY be used to augment human research, not replace it. 

By following these guidelines, you can effectively leverage AI to enhance your research process, save time, and improve the quality of your findings. 

Analyze Literature Responsibly

Generative chatbots (like basic Gemini or ChatGPT) are not academic research engines. Because they are optimized for linguistic fluency over truth, they cannot reliably retrieve primary literature and often generate highly plausible, but completely fake, journal articles and DOIs (hallucinations).

Furthermore, because AI outputs are non-deterministic, running a query twice will yield different results. This presents severe risks for systematic or scoping reviews where search replicability is required.
To protect research integrity, faculty and students must rely on academic-specific AI engines trained strictly on vetted, peer-reviewed databases.

Support Student Researchers

1. Assign Authorship and AI, in Effective and Responsible Use of AI in Research, from the University of Washington Graduate School. 

2. Direct students to use research-specific tools rather than general chat interfaces, such as: 

   • Gemini, using its Deep Research tool. 

   • NotebookLM (a closed system that uses only the materials you add, learn more here.

   • Elicit (a research assistant that automates parts of your workflow, video playlist).

   • Consensus (an academic research tool that limits its data search to the 200M published papers in Semantic Scholar, video playlist).

   • Open Evidence (a medical information platform powered by JAMA and many other medical association databases, watch Open Evidence vs. ChatGPT: A Doctor’s Perspective, 6:40).

3. Help students identify and use AI for data analysis.
   To analyze your research data effectively with AI, you must follow a structured process that protects data privacy and ensures mathematical accuracy. AIt requires specific prompting to avoid hallucinations.

   • Choose the Right Tool for the Data Type

     1. For Quantitative Data (Numbers, Stats, Surveys): Use AI models with built-in code execution capabilities (like Advanced Data Analysis features in Gemini). These tools write and run Python code in the background to calculate exact stats rather than guessing the math.

     2. For Qualitative Data (Interviews, Text, Transcripts): Use large-context models (like Claude 3.5 Sonnet or Gemini 1.5 Pro) that can process hundreds of pages of text at once without forgetting details.

   • Prepare the Data

     1. Remove personally identifiable info (names, emails, specific locations).

     2. Convert your files into clean formats. Use CSV or Excel for numbers, and TXT or Markdown for interviews.

     3. Ensure your spreadsheets have clear, single-row column headers (e.g., Participant_ID, Age, Q1_Response).

   • Apply Prompting Frameworks for Analysis

     1. For Quantitative (Statistical) Analysis: 

        • Do NOT ask the AI to "look at the data and tell you what it means." Instead, force it to use code to generate precise statistics.

        • Prompt Example (Using CREATE):

          • Context: I have a CSV file containing survey data from 200 participants evaluating a new educational app.

          • Role: Act as a Senior Data Scientist and Academic Researcher.

          • Action: Write and execute Python code to calculate the descriptive statistics for columns B through F. Then, run a chi-square test to see if satisfaction scores differ significantly by age group.

          • Target: Provide the final statistical values (p-values, means, standard deviations) in a clean Markdown table, followed by a 3-sentence summary of the main takeaway.

          • Evaluate: Ensure you rely strictly on the executed Python code outputs. Do not approximate or guess any numbers.

     2. For Qualitative (Thematic) Analysis
        AI is incredibly fast at coding transcripts and finding underlying themes. Use a multi-step prompting approach to avoid bias.

     3. Prompt Example (Using RISEN):

        • Role: Act as a Qualitative Research Specialist expert in Thematic Analysis.

        • Instruction: Analyze the attached interview transcript. Identify the top 5 recurring themes regarding user frustration.

        • Steps:

          • Read the entire text.

          • Generate a preliminary codebook with definitions.

          • Extract exactly 2 direct, verbatim quotes to justify every theme you identify.

        • End Goal: A structured thematic report.

        • Narrowing: Do not summarize the interviews generally. Focus only on user frustration. Do not paraphrase the direct quotes.

   • Verify and Fact-Check the Output

     1. Spot-Check Quotes: Always press Ctrl+F (or Cmd+F) in your original transcript to ensure the quotes the AI provided are perfectly accurate and not fabricated.

     2. Verify Code: If the AI ran Python code to get a statistical result, review the code block to ensure it targeted the correct columns and didn't ignore missing data (NaN values) incorrectly.

     3. Watch for Hallucinations: AI can confidently state a correlation exists when it does not. Treat AI insights as a "first draft" of your analysis.

Resources

• Mollick, E. (2024, May 26). Four Singularities for Research: The rise of AI is creating both crisis and opportunity. One Useful Thing. Substack.

• Tay, A. (2025, Feb 20). The Rise of Agent-Based Deep Research. Aaron Tay's Musings about Librarianship. Substack.

• Tay, A. (2025, May 11). Ai2 Paper Finder and Futurehouse PaperQA2: More transparent Deep Search for Scholars? Aaron Tay's Musings about Librarianship. Substack.

• Nuance Behavior (2025, Apr 8). Deep Research Explained: Using AI Tools for Research.

• Nuance Behavior (2025, Apr 9). AI Tools for Literature Reviews Benchmarking Report.

Overview

Much like the impact social media had on education, ignoring AI could lead to missed opportunities. On the podcast What Are the Benefits of Using AI as a Writing Partner? (Heinemann, 2025 (10:14)), two teachers discuss how they learned to integrate AI into their classrooms, making it a valuable resource for their writing workshop approach. They developed strategies for:

• guiding students to use it effectively and responsibly

• helping students overcome writer's block and enhance their creativity

• learning to ask the right questions to critically analyze AI-generated content

Writing as a Process

Writing is not just a method of turning in a grade, it is a cognitive process that helps students organize thoughts, build logical reasoning, and discover their personal voice. Their voice and the cognitive friction that comes from writing are important:

• The Homogenization Risk: LLMs generate statistically "average" writing, which can homogenize and strip away unique student voices and cultural backgrounds.

• Preserving Productive Struggle: By outsourcing writing tasks entirely, students bypass the cognitive struggle required to master critical thinking and synthesis.

• Scaffolding as an Ally: When AI is used as a diagnostic and editing assistant, it can serve as an effective, supportive writing partner.

Supporting Student Writers

Ensure students use vetted, privacy-conscious tools for writing and drafting:

• Write Right (Google Doc) lists free online writing tools to help analyze, revise, and improve writing. The free Grammarly Chrome add-on works in Brightspace, too. 

• Turnitin Draft Coach: Have students install the Turnitin Draft Coach Google Docs extension. It allows them to check and re-check their drafts for grammar, citation formatting, and similarity matches while writing.

• Gemini Access: Remind students that as a Google School, CSS provides free, secure access to Gemini for writing feedback.

How to: Scaffold Writing & Prevent Total AI Outsourcing

• Version Tracking: Require students to use Google Docs and add you as an editor so you can observe their writing process. You can view the document's Version History (under File) to see what they wrote at any given time.

• Structure Checkpoints: Break major papers into progressive phases (proposals, outlines, annotated bibliographies, rough drafts, peer reviews, and final revisions). This models writing center best practices and makes copying-and-pasting the final product from AI highly tedious.

• The AI Appendix: Require students who use AI for drafting or refining to attach an "AI Log" to their paper, disclosing the prompts entered and writing a short paragraph explaining which AI edits they chose to accept or reject.

• Flipped Drafting: Have students draft a research outline by hand in class. For homework, have them ask ChatGPT for an alternative outline and write a short comparison explaining why their human outline is more rhetorically effective than the machine's statistical average.

Resources

• AI and Writing Assignments (4:18). Analyzes the benefits and challenges GenAI presents in writing assignments, and discusses how proper guidance and assignment construction can develop key writing skills.

• AI’s Use in Structuring Writing Assignments (2:38). Explains how to use generative AI to design effective writing assignments that align with course objectives. An example prompt that can be adapted to any discipline is provided.

• AI for Helping Students Revise Writing Assignments (2:02). Outlines how instructors can use GenAI to aid students in the writing revision process with a sample prompt and best practices.

• Critical Scholarly Analysis for ChatGPT is a custom bot that acts as an expert reviewer for academic papers, providing critical, scholarly analysis. Share this with students.

• Quick Start Guide to AI and Writing (08/2023). MLA-CCCC Joint Task Force on Writing and AI.