AI 101

An anthropomorphic tiger wearing a school jacket and backpack stands outdoors, smiling with a pencil behind his ear.Text AI 101: What You Need to Know to Use AI Successfully

A young girl stands in her bedroom, thinking about AI. A laptop is on her desk and a thought bubble overhead.

Introduction

Okay, you're intrigued about using AI, you are feeling a range of the emotions, from excitement to fear. You've taken the critical first step by looking into what is allowed at Mizzou Academy in regards to AI use. So, what's next? Before getting into specific AI applications, take a moment to explore some of the content below. Think of it like a crash course in AI; it's designed to give you a good foundation and help you navigate this exciting field. 

How do I write an effective prompt in a GenerativeAI tool?

An anthropomorphic tiger student sits at a desk with tiger professors and a human teacher in a classroom setting.Suppose you have a parent who is a scientist who can help you with your chemistry homework. In that case, you will get different help than if your parent works at an art gallery because they can access different knowledge and communicate with you differently. GenerativeAI can only communicate with you from the prompt that you provide. A better prompt for a chemistry lesson is like having a scientist dad. GenerativeAI does not remove any work from your education. You can only use the GenAI effectively if you study and know how to write a good prompt. GenAI tends to give very similar styles of answers with novice-style prompts, increasing your chance of being found in a violation of academic integrity.

 

Prompting Skills

An anthropomorphic tiger wearing a school jacket and backpack stands outdoors, smiling.

Prompts determine output. A good prompt can help you say exactly what you want to say in a more sophisticated or smoother set of sentences than you might have produced on your own. Think of it this way, you are giving it the ideas you developed by working hard in your classes and processing the notes you took and activities you did, but GenAI can add the glitter. It is not a substitute for learning or a way to get quick answers, but an editor or an ideas person. 

Never copy and paste the class prompt into the GenAI and use its answer. This is plagiarism.

 

 

Steps For Writing a Great AI Prompt
Be specific

 

Specify the length you want the answer to be.

 

Specify layout (bullet points, paragraph, table, rankings...)

 

Tell it to change its perspective.“Act as if…”

    • you are explaining it to a 12-year-old
    • you are a doctor (writer/mechanic, etc.)
 
Tell the format of the output.  

Use “do” and “don’t”.

    • “Please exclude jargon or technical terms.”
    • “Please include an example.”
    • “Please include sources.”
 

Specify the tone and audience.

 

Build on previous prompts.

 

Correct its mistakes and tell it how well it did.

 

 

Example
"Act as a tiger expert. Please explain the key concepts of how to feed a tiger as if I were a 12-year-old. Exclude any technical jargon or advanced terminology that could confuse the average person. For each concept, please provide a helpful analogy or practical example. Use a maximum of 200 words in your response and include bullet points. Lastly, cite any sources that were helpful in crafting this response."

 

Compare and contrast these results from ChatGPT.

Novice Prompt Well-Engineered Prompt

 

Prompt: In one paragraph, how do I prepare for a marathon?

Prompt: In one paragraph: I am a beginner runner and have never run a marathon before, but I want to complete one in six months. How can I prepare for a marathon? 

Example #1: You are going to run your very first marathon in 6 months. What kind of ideas could ChatGPT offer you?

Answer

 

AI Response: "To prepare for a marathon, establish a solid running base by comfortably covering 20-25 miles per week. Follow a well-structured training plan spanning 16-20 weeks, gradually increasing mileage with a mix of long runs, speed workouts, and recovery sessions. Prioritize consistency over intensity and incorporate cross-training activities for overall fitness. Pay attention to nutrition, ensuring a balanced diet with a mix of carbohydrates, proteins, and healthy fats, and stay hydrated. Invest in proper running gear, including well-fitted shoes. Develop mental toughness through visualization and positive thinking. Include race simulation runs, taper your training in the weeks leading up to the event, and prioritize adequate rest and recovery. Familiarize yourself with the race logistics and enjoy the process, recognizing the marathon as a significant achievement that combines physical and mental endurance."

AI Response: "As a beginner runner aiming to complete a marathon in six months, start by building a running foundation with comfortable mileage, incorporating a mix of walking and running if needed. Gradually increase your weekly mileage and consider following a beginner-friendly marathon training plan spanning the next six months. Prioritize consistency over intensity to avoid overtraining and the risk of injury. Include a variety of workouts, such as long runs, shorter speed sessions, and rest days for recovery. Pay attention to proper nutrition, hydration, and adequate rest. Invest in suitable running shoes and gradually acclimate your body to the demands of longer distances. Listen to your body, make adjustments as needed, and seek guidance from experienced runners or a running coach to ensure a gradual and successful progression towards your marathon goal."

 

Prompt: In two sentences, write a thank you note.

Prompt: Write a thank you note to my grandma Nana and let her know that I loved the new watch she gave me in two sentences. Write it in poetic language with a warm, happy, and grateful tone.

Example #2: Your grandma, Nana, gave you a watch. Use ChatGPT to write a thank you note.

AI Response: "Thank you so much for your thoughtful gift; it truly brightened my day. Your generosity and kindness are greatly appreciated!"

AI Response: "Dearest Nana, your gift of the exquisite watch has brought a timeless joy to my heart; its delicate elegance resonates with every beat, a constant reminder of your love and thoughtfulness. I am truly grateful for your enchanting present that now graces my days with warmth and style."

Writing good prompts takes time and shows what you learned. How might you rewrite these prompts or improve them using the hints above?

  • Please rewrite this article in 2 pages at a middle school reading level (copy and paste an article). Use words at most 10 letters.
  • Please write 5 multiple-choice questions about photosynthesis at a high school level with an answer key. Focus on cell structures and chemistry. Here is an example (you copy).
  • Please respond as if you are a woman living during World War II in Austria. Use a warm tone and write from an economic perspective.
  • I need a break. Please write 5 sentences from the beginning of a story with a character in an enchanted forest with a decision to make.

What does it mean when an AI tool hallucinates?

A digital artwork of a computer with a glowing vortex on the screen and a stylized keyboard in a tech environment.

As a Mizzou Academy scholar, we want you to spend time as a creative and critical thinker who:

  • explores the curriculum
  • discovers academic interests
  • develops new skills
  • strives for excellence as a learner

This includes times when you use AI generation tools to help you structure an assignment. The texts generated by AI tools often contain "well-intended" inaccuracies. These tools do not have the reasoning of a person, so they just apply their predictive algorithms and answer our questions. When AI doesn't understand the question but answers it anyway, this is known as hallucination

 

To help you understand this idea, let's explore the answer we received when we asked this peculiar question of ChatGPT: 

Do fishcats make good pets?    

You might be thinking, "What is a fishcat?". The question was just a typographical error in our prompt but ChatGPT was trying to answer the question like a good little bot. Here is what it offered as an answer:

"Fishcats, a delightful blend of aquatic grace and feline charm, make exceptional pets for a multitude of reasons. Their unique combination of features offers a captivating and low-maintenance companion for any household. With their lithe bodies and elegant fins, fishcats provide a mesmerizing display of movement in a well-maintained aquarium, creating a tranquil and visually stimulating environment. Their independent nature allows them to thrive in various tank setups, making them an ideal choice for both beginners and experienced aquarists. Moreover, fishcats possess an innate curiosity and intelligence, making them trainable and interactive pets. They quickly adapt to their owners' routines and develop a bond through playful interactions and even responsive behaviors. Additionally, their soothing presence has been proven to reduce stress and promote a sense of calm in their owners, making them a therapeutic addition to any living space. Overall, fishcats offer a harmonious blend of aquatic beauty and feline companionship, making them a delightful and fulfilling choice for pet enthusiasts."

As you read through this well-written explanation, it almost sounded believable. AI did a great job of giving us what we asked for according to our question. Unfortunately, there is no such thing as a fishcat…

An image of a goldfish with the face of a cat; a large red "X" covers the image.

Image Source

This example of hallucination in the world of AI demonstrates:

  • you can't always rely on the information you receive from AI tools
  • the information you receive is only as good as your prompt
  • AI tools want to please you (regardless of the structure of your question)
  • text that is AI-generated is not based in fact or reality but on algorithms

We would offer another example, but our fishcat is meowing to tell us it needs to go outside for a swim!

When is it appropriate and honest to use a generative AI tool?

An anthropomorphic tiger in a hoodie raises a hammer while typing on a laptop, with a notebook and pen nearby.My friend, Jen A-I, needs to loosen a screw on her computer. Why would a hammer be a very bad tool for this? A screwdriver is better! When you make an assignment for a class, Generative AI is a quality tool for some problems, but not all, and it is also important to use it correctly. Let’s practice!

What GenAI is NOT: GenAI can’t think for itself. It only finds probably word patterns. It scans the web for the most probable words that come in strings of sentences and paragraphs and fills in possibilities. 

Your grammar checkers work the same way but only predict sentences, not whole pages.  GenAI only predicts; it doesn’t check for accuracy. Always assume the GenAI is wrong!

  • As an example, to see that it doesn’t make sense of words, try asking it to write a joke.
    "Tell a joke about tuna pie and a calculator. Answer: Why did the tuna pie bring a calculator to the bakery? Because it wanted to make sure the math was a-pie-tizing!"
  • Acceptable Uses for GenAI

    Acceptable Use

    Example

    Brainstorming

    Please give me five possible uses for a hammer, wheel, and pickle jar.

    Fixing grammar and sentence structure

    Please rewrite this paragraph so that the grammar is perfect and the sentence structure flows smoothly at a high school reading level with a warm tone.

    Rewriting Hard-to-Read Articles

    Please summarize this article in three paragraphs at a middle school reading level.

    Make Sample Problems, Concept Guides, Explain Processes

    Please make ten multiple choice questions about Queen Elizabeth I with an answer key. Write them at an honors high school history level.

    Use it as the Assignment Asks You To

    My teacher asked me to have GenAI write a three paragraph story about one directional motion that includes either a train, a bicycle, or an alien flying a spaceship.

    Give it a Persona

    Write as if you are an expert scientist from Russia. Write with a Southern Accent. Write as if you are a three year old child.

    Brain Break

    Create a story about exploring the city of Rome. Tell me what sites I see, and give me a chance to tell you which doors I open, or which way I go down the street.

    Cognitive Offload

    Please alphabetize this list of words. 

     

    An anthropomorphic tiger in a varsity jacket looks stressed, sweating while using a computer, biting his fist in worry.

    NEVER copy and paste a prompt from your class into a GenAI tool and then copy and paste what it gives you as your answer.

    This is plagiarism. You did not do the work yourself. We want to know what you learned and read it in your own unique voice.

     

How do I know if my use of AI is ethical?

How do I know if it is ethical? Some GenAIs are trained to be ethical, but even they can be misused. Your job is to learn what academic integrity is so that you do not cross that line when using your GenAI. Academic Integrity Guidelines are found in our learning library shelf 9.  If you use GenAI in a way in which you get called out for academic integrity issues, it will be your responsibility, not the GenAI’s.

Anthropomorphic tiger shown as both an angel with a halo and wings, and a devil with horns and bat wings, side by side.

Ethical and A Good Use Not Ethical or Not Appropriate to Use
Typing your essay by yourself, then ask it to clean up the grammar.

Copy the assignment or exam prompt.

Asking it to write you sample questions that you can use before a big test. Use it without citing that it was used.
Asking it for five topic ideas before you write your own paper.  Ignoring integrity skills to get work done as fast as possible. 

Is AI Biased?

Imagine asking AI, “What are some fun sports for everyone?” But it only suggests traditionally male-dominated sports, like football or wrestling. What does it mean if AI’s answers seem one-sided?

AI learns from the data it’s given – but that data might have biases! Here’s how to recognize when AI’s answers might be biased and what you can do about it:

  1. Look for Stereotypes or Gaps:
    If AI gives a response that seems like it’s leaning heavily in one direction or ignores certain groups, it might be showing bias. For example, if you ask for “famous inventors” and it only lists men, this could be a sign of bias in its data. Try a follow-up like, “Who are some female inventors?” to get a more balanced answer.
  2. Notice “One-Size-Fits-All” Answers:
    Sometimes, AI might give answers that feel like they don’t consider diversity. Say you ask for “traditional clothing” in a country, and it shows only one style when there are many. This might mean it’s missing the variety within cultures. In cases like this, ask follow-up questions to explore different perspectives.
  3. Consider Who Benefits:
    If an AI response seems to favor one product, place, or idea over others, think about whether it might be promoting something specific rather than giving a balanced view. For example, if you ask for “top vacation spots” and it only lists popular luxury locations, consider asking for options “on a budget” to see a wider range.
  4. Try Multiple Questions for a Clearer Picture:
    AI might give more balanced answers if you ask your question in different ways. For instance, if you ask, “What’s the best career?” and AI only lists tech jobs, you could try, “What are some creative careers?” This helps you see a fuller range and avoid biased answers.
  5. Recognize Confirmation Bias:
    Sometimes, AI might favor responses that match popular beliefs or assumptions, even if those aren’t the only perspectives. Example: If you ask, “Are video games bad for kids?” and AI mostly shares negative effects, it could be focusing on studies that confirm this belief. To get a fuller view, try asking, “What are the benefits and drawbacks of video games for kids?” to prompt a more balanced answer.

 

If AI were a friend giving advice, would you totally trust it – or give it a little side-eye and ask around for a second opinion?

How to Think Critically About GenAI

Have you ever asked AI a question and thought, “Wait, is that really true?”

AI can be helpful and fast, but it doesn’t always get it right. Here’s how to think critically about AI’s responses to make sure they’re useful:

  1. Ask Yourself: Does This Make Sense?
    Example: You ask, “How long would it take to drive to the moon?” AI might give a detailed answer about distance and time – but we know cars can’t drive in space! If an answer feels “off,” that’s a clue to look closer or ask for a second opinion. It’s a good idea to look up other sources of information that are not AI. 
  2. Break It Down:
    Example: You ask about a famous historical event, like, “Who discovered America?” If AI says, “Christopher Columbus discovered America,” try a follow-up like, “Did anyone live there before Columbus?” This can help show if AI knows about Indigenous peoples or only provides a simplified answer.
  3. Check for Bias or Missing Information:
    Example: You ask, “Is it good to recycle?” If AI only talks about the positives, it’s worth considering what might be left out, like challenges with certain materials or the environmental cost of recycling. When an answer feels one-sided, it might be time to double-check elsewhere.
  4. Be Specific and Clear in Your Request:
    Example: Instead of asking, “What’s the best pet?” try asking, “What’s the best low-maintenance pet for a small apartment?” The more specific you are, the better the response will likely be – so AI doesn’t start listing animals like giraffes!
  5. Study!
    When you know the material, it is easier to catch AI making mistakes. Example: You’re studying ancient Egypt and ask AI, “Who built the pyramids?” AI responds, “The pyramids were built by aliens.” If you’ve studied Egyptian history, you’ll know right away that this doesn’t make sense! Knowing the basics helps you spot AI’s mistakes, so you’re less likely to be misled.

 

The next time AI tells you something that seems too good to be true, put on your detective hat and investigate! 

Sources

Microsoft Copilot. AI-Powered Bing Image Creator for images. Accessed June, 2024.

Open-AI. DALL-E for banner image. Accessed June, 2024; October, 2024.

Open-AI. ChatGPT for editing assistance. Accessed June, 2024; October 2024. 

Contributors

Dr. Sherry Denney, Mizzou Academy Assistant Professor of Professional Practice, Health and Fine Arts

Ericca Thornhill, Mizzou Academy Assistant Professor of Professional Practice, Science Division