1 of 19

INTRODUCTION TO USING GEN AI: BASICS OF PROMPTING WITH GEMINI

Matt Kilgas

Associate Professor, School of Health and Human Performance

Teaching and Learning Scholar, Center for Teaching and Learning

​

Scott Smith

Instructional Technologist, Center for teaching and Learning

2 of 19

Learning Objectives

  • Explain what generative AI is and the basic mechanics of how it functions.

​

  • Identify the benefits and limitations of Gen AI

​

  • Apply prompting techniques to complete basic tasks with Gen AI.

2

3 of 19

Outline

  • Activity: Vibe Check-Where are We Now?

​

  • Activity: Accessing Google Gemini

​

  • Presentation: AI and LLMs (Mechanics & Key Points)

​

  • Activity: Continuing the Conversation

​

  • Activity: Examining Output

​

  • Presentation: Prompting Basics (The Framework)

​

  • Presentation: Improving the Prompt (Meta-Prompting)

​

  • Activity: Improving the Prompt

​

  • Presentation: University Policies & Data Safety

​

  • Presentation: NMU Google Education Agreement

​

  • Activity: Completing a Task with AI

​

  • Activity: Vibe Check-Where are We Now?

3

4 of 19

Vibe Check-Where are We Now?

  • Think-Pair-Share
    • Why are you here? Why did you sign-up for this workshop?

​

    • What are your main questions/concerns when it comes to AI?

​

    • Have you used AI before? If so what for?

​

    • What have you heard about AI (the good, bad and ugly)?

4

5 of 19

Activity 1 – Accessing Google Gemini

    • Visit Google Gemini
      1. Log in using your NMU ID and Password

​

    • In the “Ask Gemini” text window, write “Explain Generative AI to me in simple and non-technical terms”

​

    • Hit Enter

​

5

6 of 19

AI and LLMs

  • Generative AI (GenAI): Technology that can create original text, images, video and other content that mimic human work.

​

  • Machine Learning: Algorithms that can “learn” the patterns of training data and, subsequently, make accurate inferences about new data.

​

  • Deep learning: Subset of machine learning driven by multilayered neural networks whose design is inspired by the structure of the human brain.
    • Neural Networks: Interconnected layers of mathematical "nodes" that pass data along, adjusting weights to recognize complex patterns.

​

  • Large Language Models (LLMs): Models capable of understanding and generating natural language and other types of content.
    • LLMs work as statistical prediction machines that repeatedly predict the next word in a sequence. They learn patterns in their text and generate language that follows those patterns.

​

6

7 of 19

Activity 2: Continuing the Conversation

    • In the same window as before (under the output), ask a follow up question about the output including a new definition from the handout
      1. For example: What is a token? How is it used in Neural Networks?

​

    • Hit Enter

​

7

8 of 19

Key Point about AI

  • AI is a prediction machine, it does not think, understand, or know things.

​

  • AI can produce polished outputs that are wrong, incomplete, biased, or inappropriate for the context.

​

​

8

9 of 19

Activity 3: Examining Output

    • In a new window, ask the AI to explain a concept from your class, or a topic you have expertise in.

​

    • Hit Enter

​

    • What did it do well? What did it miss?

​

9

10 of 19

Prompting Basics

  • The prompt should include:

​

    • Role: Act as a [Role].
    • Context: I am writing a [Task] for [Audience].
    • Instructions: Tone, style, and length.
    • Constraints: Guardrails and specific background.
    • Examples and knowledge bases: Information the AI should use in the output
    • Summary Command: 'Now, write this...'

​

Role:�Act as an experienced finance teacher and curriculum designer.

Context:�I am writing a 60-minute lesson plan for university freshman who are learning to create and manage a personal budget for the first time. The students have little prior knowledge of budgeting, so the lesson should introduce the concepts in a simple, engaging, and practical way.

Instructions:�Organize the lesson using the following sections: Learning Objectives, Materials, Introduction, Direct Instruction, Guided Practice, Independent Practice, Assessment, and Closure. Include one real-world example of a monthly budget for a young adult. Incorporate at least one interactive activity where students create their own budget using a realistic scenario.

Constraints:�The lesson should be completed in approximately 60 minutes.

Examples and Knowledge Base:�Use evidence-based instructional practices such as active learning. Base budgeting examples on realistic income and expenses for a part-time job or an entry-level position. The lesson plan should align with common personal finance learning standards.

Summary Command:�Write a draft of the lesson plan.

​

10

11 of 19

Improving the Prompt

  • It won’t be perfect on the first try.
    • Treat the first output as a rough draft. Explain what the output did well and what it didn’t do well.

​

  • Meta-Prompting:
    • What questions do you have about the task?
    • What info do you need from me to make this prompt effective?
    • You are an expert prompt engineer, write a prompt for…

​

11

12 of 19

Activity 4: Improving the prompt

    • In the same window as Activity 3 (under the output), use the prompting information to ask the AI to explain a concept from your class, or a topic you have expertise in.

​

    • Hit Enter

​

    • What changed from the last output?

12

13 of 19

University Policies

​

13

14 of 19

University Policies: Key Points

  • Do not provide information containing:
    • Personally Identifiable Information:
      • Student or employee information: names, NMU ID numbers, birthdates, addresses, grades, etc.

​

    • Confidential Data: Non-public financial reports, proprietary research, or information protected by copyright, NDAs, or that is private/confidential.
      • Unless you have written permission to do so.

​

    • Financial & Payment Card Data: Credit card numbers, banking details, cardholder data, or financial transaction logs.

​

14

15 of 19

NMU Google Education Agreement

  • NMU has licensed Google Gemini for university work.

​

  • This agreement prevents Google from using our inputs to train their models.
    • Fully compliant with FERPA, COPPA, and HIPAA.

​

  • If it is appropriate for you to save data or file to your institutional Google Drive, then it is appropriate for Gemini.

​

15

16 of 19

Activity 5: Completing a task with AI

    • Open a new chat:

​

    • Ask AI to complete a task you find tedious
      1. Write a draft of an email
      2. Create a grocery list
      3. Brainstorm a class activity
      4. Revise instructions
      5. Organize a to-do list

​

16

17 of 19

The CARE Framework

  • AI output always needs to be reviewed and edited.

​

  • CARE Framework
    • C-Correct: Is the information accurate?

​

    • A- Appropriate: Does it fit the intended context?

​

    • R-Responsible: Double-check that you did not leave in any private, sensitive, or confidential information.

​

    • E-Edited: Never just copy-paste.

17

18 of 19

Activity: Vibe Check – Where are we now

  • Think-Pair-Share
    • Biggest take-aways
      • What did you learn?

​

    • Remaining questions

​

    • Biggest Concerns

18

19 of 19

Questions

AI Disclosure Statement: This presentation was collaboratively developed using Google Gemini. All concepts, frameworks, and policies were supplied entirely by the instructor, not independently generated by AI. The AI functioned strictly as a formatting and synthesis tool to organize these provided materials into a cohesive structure. The final presentation and accompanying materials were thoroughly reviewed, edited, and finalized by the instructor to ensure accuracy, proper attribution, and alignment with institutional goals.

19

Questions?