1 of 31

AI Prompting

For Faculty and Staff at USF

Prof. & Chair Allison Thorson, athorson@usfca.edu

April 7, 2026

2 of 31

How is AI changing the playing field?

Since access to “an answer” is now easier than ever, the smartest people in the room are now those who ask the right question(s).

3 of 31

What is prompt engineering?

Prompt engineering is the practice of designing and refining inputs (prompts) to guide AI models toward generating the most useful and accurate responses. It involves understanding how language models interpret instructions and experimenting with phrasing, context, or structure to optimize outputs. In essence, it’s about learning to “speak the model’s language” so it performs better for a given task.

4 of 31

Common Large Language Models (LLMs) & data

** Founded by former OpenAI employees

5 of 31

Things to consider…

  • Not the focus of today’s session:
    • How data is generated from prompts
    • Tokens
    • Context windows
    • Privacy (settings, confidentiality, and LLMs using your prompts to train the model)
    • Laws and AI (EU & South Korea)
    • AI and employment
    • Hallucinations
    • Frankensteining
    • Ethics
    • Bias
    • Overconfidence in outputs / over-reliance risks

Today’s focus: prompt engineering!

6 of 31

Prompt engineering techniques we’ll focus on today:

  • Zero-shot

  • One-shot

  • Few-shot

  • Chain-of-Thought (CoT)

  • Instruction-Based Prompting

  • Advanced Prompting

  • Domain-Specific Prompting

  • A/B Testing
  • Cristello, B. (2024)
    • Clarity & Specificity – avoid ambiguity.
    • Context & Background – provide relevant details.
    • Desired Format – define response structure.
    • Examples & Templates – guide AI with few-shot prompting.

  • Agility AI. (2024) - Mastering Prompt Engineering. Episode 1 [summary of several standard industry frameworks]
    • Instructions – clearly and directly tell the AI what you want it to do (be direct and to the point)
    • Information – provide context and the data the AI needs to fulfill your instructions
    • Examples – show the AI what you are looking for in terms of style, form, format
    • Output format – how should results be shown

  • Lo, L.S. (2023). The CLEAR path: A framework for enhancing information literacy through prompt engineering, The Journal of Academic Librarianship, 49 (4). https://doi.org/10.1080/10875301.2023.2227621
    • CLEAR: Concise, Logical, Explicit, Adaptive, and Reflective

Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P.,

Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-

Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D.

M., Wu, J., Winter, C., … Amodei, D. (2020). Language models are few-

shot learners. Advances in Neural Information Processing Systems, 33,

1877–1901. https://arxiv.org/abs/2005.14165 (**not entirely attributed to

one set of authors)

This field is relatively new—mainstream adoption began around 2022.

7 of 31

Zero-shot prompt – give the model a prompt with no examples

  • Translate the following English sentences into French. "The cat is sleeping on the chair.” [zero-shot]
  • Write bullet points for my resume summarizing my job as a barista at Starbucks.
  • Create a scavenger hunt for my 8-year-old son’s birthday.

8 of 31

Zero-shot

  • If I currently have a GPA of 1.90 and have completed 32 credit units, what GPA do I need for the next 16 credit units to earn a cumulative GPA of 2.0?”

9 of 31

Zero-shot

  • "Why are data centers bad for the environment?”

Data centers can be harmful to the environment primarily because they consume enormous amounts of energy and resources. These facilities run thousands of servers continuously, requiring vast amounts of electricity not only to power the machines but also to cool them and prevent overheating. Much of this electricity still comes from fossil fuels, which leads to significant greenhouse gas emissions and contributes to climate change. Additionally, data centers use large quantities of water for cooling systems, which can strain local water supplies, especially in dry regions. The production and disposal of hardware also create environmental issues, as manufacturing servers requires rare minerals and generates pollution, while outdated equipment contributes to electronic waste. Together, these factors make data centers a growing environmental concern in an increasingly digital world.

10 of 31

Zero-shot – how an instructor might use this

  • AI-generated answer to the question: "Why are data centers bad for the environment?”

  • For this short answer question, dissect this AI-generated answer. Do NOT rewrite the paragraph directly. Instead, annotate and analyze it. (a) Identify inaccuracies or misleading claims. Point out any statements that are false, exaggerated, or oversimplified. Briefly explain why each is problematic. (b) Validate correct claims. Identify statements that are accurate. Strengthen them by adding specific details, data, and properly cited sources (peer-reviewed articles, government reports, or reputable organizations). (c) Add missing context. Incorporate important concepts discussed in class that are absent (e.g., regional energy mix, efficiency improvements, renewable energy adoption, lifecycle analysis, or policy considerations). (d) Provide a reflection (150–250 words). What misconceptions or gaps might a reader have if they relied only on this AI-generated answer? How could those misunderstandings affect public opinion, policy, or decision-making?

Data centers can be harmful to the environment primarily because they consume enormous amounts of energy and resources. These facilities run thousands of servers continuously, requiring vast amounts of electricity not only to power the machines but also to cool them and prevent overheating. Much of this electricity still comes from fossil fuels, which leads to significant greenhouse gas emissions and contributes to climate change. Additionally, data centers use large quantities of water for cooling systems, which can strain local water supplies, especially in dry regions. The production and disposal of hardware also create environmental issues, as manufacturing servers requires rare minerals and generates pollution, while outdated equipment contributes to electronic waste. Together, these factors make data centers a growing environmental concern in an increasingly digital world.

11 of 31

12 of 31

One-shot prompt – give the model a prompt with one example

  • Attached is an example of a recent speech Beto O'Rourke gave on YouTube (Example here: https://youtu.be/YioVe6RlkOw?si=bgVPKgG20LabdIWY). After learning his speech style, act as a speech writer for Beto O'Rourke and create a speech for him, addressing the redistricting of house seats in Texas.

  • Attached is an example of a department’s Winter newsletter that was sent to current students, alumni, faculty, and friends. Using the same content and voice, edit this newsletter – tailoring it to potential students and parents.

13 of 31

One-shot in a series of prompts

  • ** remove any student identifying information (from past and current letter)

  • Attached is a letter of recommendation that I have written for students applying to Leo T. McCarthy Center for Public Service and the Common Good, USF in DC program. Using the same writing style, tone, and voice, generate a very strong – top tier - letter of recommendation for Jane Doe. Incorporate specific information about Jane which I will include in subsequent prompts.

    • Mention that I have known Jane for approximately 1 year, as she was a student in my COMS 203, Communication and Everyday Life course.
    • When writing this letter highlight the reasons why Jane is a good fit for this program, including: She has a proven record of turning advocacy into action—whether through legislation, community program design, or statewide youth organizing. Mention her community and advocacy work which includes: Hope Means Nevada (Teen Committee Co-Chair & Founder of Northern NV Chapter); Expanded youth mental health advocacy statewide by partnering with policymakers on Senate Bill 249b, delivering speeches to large audiences, and organizing statewide youth summits;
    • Mention her work with the Domestic Violence Resource Center (Youth Program Development Intern) where she designed a rotating four-week curriculum for shelter youth, built partnerships with local nonprofits, and implemented trauma-informed onboarding systems. Laudato Si’ Movement of California (Project Contributor):
    • Share that she regularly speaks on youth mental health, policy reform, and community resilience; collaborated on school district mental health curriculum.
    • Incorporate the following experiences Jane has had into this letter: she is a Sophomore at USF, Communication Studies major with minors in Public Policy & Urban and Public Affairs, a Getty Honors College Fellow, Member of the USF women’s soccer team, public safety officer and social media manager
    • Provide this as an unformatted text document that can be easily copied and pasted.

14 of 31

One-shot in a series of prompts

  • ** remove any student identifying information (from past and current letter)

  • Attached is a letter of recommendation that I have written for students applying to Leo T. McCarthy Center for Public Service and the Common Good, USF in DC program. Using the same writing style, tone, and voice, generate a very strong – top tier - letter of recommendation for Jane Doe. Incorporate specific information about Jane which I will include in subsequent prompts.

    • Mention that I have known Jane for approximately 1 year, as she was a student in my COMS 203, Communication and Everyday Life course.
    • When writing this letter highlight the reasons why Jane is a good fit for this program, including: She has a proven record of turning advocacy into action—whether through legislation, community program design, or statewide youth organizing. Mention her community and advocacy work which includes: Hope Means Nevada (Teen Committee Co-Chair & Founder of Northern NV Chapter); Expanded youth mental health advocacy statewide by partnering with policymakers on Senate Bill 249b, delivering speeches to large audiences, and organizing statewide youth summits;
    • Mention her work with the Domestic Violence Resource Center (Youth Program Development Intern) where she designed a rotating four-week curriculum for shelter youth, built partnerships with local nonprofits, and implemented trauma-informed onboarding systems. Laudato Si’ Movement of California (Project Contributor):
    • Share that she regularly speaks on youth mental health, policy reform, and community resilience; collaborated on school district mental health curriculum.
    • Incorporate the following experiences Jane has had into this letter: she is a Sophomore at USF, Communication Studies major with minors in Public Policy & Urban and Public Affairs, a Getty Honors College Fellow, Member of the USF women’s soccer team, public safety officer and social media manager
    • Provide this as an unformatted text document that can be easily copied and pasted.

Rough draft

15 of 31

16 of 31

Few-shot prompt – give a few examples of what you want before you ask your prompt (set the tone with the style you want)�

  • Act as a party planner and create a Minecraft scavenger hunt for my 8-year-old son's backyard birthday. Make each clue sound like Steve (Jack Black's) character. Bring in catch phrases from the movie such as: Example 1: "I ... am Steve". Example 2: “Chicken Jockey!” Example 3: “First we mine, then we craft… LET’S MINECRAFT!” Example 4: Steve's Lava Chicken. Hide clues under the swing, on the trampoline, under a chair cushion, and in the bird feeder.

  • Act as a hallmark card maker and create silly Valentine’s Day sayings for cards based on animals. Use a play on words like: Example 1: Wallaby my valentine. Example 2: I goat crazy for you. Example 3: a skunk with the saying "I stink I love you"; Example 4: "Piranha be my valentine?” Example 5: You are Otter(ly) adorable.

17 of 31

Few-shot

  • Act as chair of the Department of Communication Studies. Create catchy names for events during the Spring semester 2026 that are consistent with previous events: Example 1: "Pizza with the Profs". Example 2: Pie Day with Lambda Pi Eta. Example 3: COMS Game Day (Day at the ballpark). Name events in a way that rhymes or uses alliteration.

18 of 31

How a student could use few-shot prompting…

  • I am second semester college Psychology student writing a literature review. My assignment is copied and pasted here: … Attached are 3 sample literature reviews provided by my professor written by other students. Analyze these papers for structure, format, and assignment requirements. Last, create a literature review with peer-reviewed citations on oppositional defiance disorder in adolescence.

How might we need to change our assignments to adapt to this new technology?

19 of 31

20 of 31

CLEAR framework, Lo (2023)

  • Concise: use clear and concise language. (understanding tokens)
    • USE “Create” rather than “Can you provide me with a detailed explanation of…”
  • Logical: a logically structured prompt helps the AI model understand the context and relationships between various concepts.
    • “List, describe, include” – add steps if necessary
  • Explicit: provide precise instructions about the format, content, or scope of your desired output.
    • Output should be in … format. OR “identify 5 elements and summarize each in 100 words or less”
  • Adaptive: after your initial prompt, review the content for additional keywords, context, or parameters.
  • Reflective: think critically about the AI tool's answer. Does it make sense? Are there perspectives or experiences that are missing?

21 of 31

Act as a college professor who needs to create ADA-compliant image descriptions (alt text) for the visually impaired using the image attached. Keep descriptions short. Focus on the information the image is meant to convey within the context of the page. Describe what is physically present. Describe the image based on why it is there. Include any text in the description. Present output as unformatted text that can be easily copied and pasted.

  • Adaptive: after your initial prompt, review the content for additional keywords, context, or parameters

22 of 31

Reflective: think critically about the AI tool's answer. Does it make sense? Are there perspectives or experiences that are missing?

  • Image description:�The image shows the front cover of a book titled “Generative AI Professional Prompt Engineering Guide.”�Beneath the title, a smaller subtitle reads:�“Prompt Engineering Excellence: Unlock the Full Potential of Generative AI.”�At the bottom right corner, the author’s name, George Tome, is displayed.
    • The main visual is a striking split-face illustration. The left half depicts a human head rendered in soft white and gray tones, with a realistic brain visible inside a translucent skull. This side represents organic intelligence and creativity.
    • The right half of the face transitions seamlessly into a robotic, mechanical head filled with glowing circuitry, wires, gears, and illuminated components. Blue and orange lights glow within this metallic structure, suggesting artificial intelligence and advanced technology.
    • A bright vertical beam of light runs down the center, dividing the human and machine halves, symbolizing the fusion of human thought and AI computation. The background is dark and futuristic, with subtle glowing orbs and digital patterns that emphasize the theme of technology and intelligence.

23 of 31

24 of 31

More advanced techniques – building your own tools:

  • Prompt 1:

“Act as an internship class instructor for a college course. I want to use CHATGPT to act as a mock interviewer. I want CHATGPT to ask an interview question and have students answer this question by clicking on the microphone in the chat. Based on students' responses (a weak answer or a strong answer) adjust what the next follow-up question will be as if you were having a conversation with them. “

  • Prompt 2:

“One thing I would like you to add to this assignment is - after asking 7 questions, provide feedback to the student on the strengths of their answers, weaknesses, and suggestions on how to improve. A strong answer will use the STAR method (a structured technique for answering behavioral interview questions that involves the answer to ‘Tell me about a time when...’ detailing a specific Situation, Task, Action, and Result)”

25 of 31

Output: (Instructions for students: Copy and paste this prompt into an LLM and respond verbally.)

26 of 31

27 of 31

Some things to notice:

  • Public vs. private - No identifying information should be shared (name, student IDs, university name, class name, link to personal course or instructor information, etc.)

  • Let your question guide the prompt technique – do not let the technique guide the question

28 of 31

It’s all about communicating effectively with the LLM.�

  • Keep a “human-in-the-loop” - put yourself into the process
  • Break down complex tasks into manageable chunks (Prompt 1: First…. Prompt 2: Now that we know A, now ….)
  • Reflect on the output and think critically
  • Refine prompts and the output – this is an iterative process
  • Experiment and get creative

29 of 31

Prompt engineering techniques from today:

  • Zero-shot

  • One-shot

  • Few-shot

  • Chain-of-Thought (CoT)

  • Instruction-Based Prompting

  • Advanced Prompting

  • Domain-Specific Prompting

  • A/B Testing
  • Cristello, B. (2024)
    • Clarity & Specificity – avoid ambiguity.
    • Context & Background – provide relevant details.
    • Desired Format – define response structure.
    • Examples & Templates – guide AI with few-shot prompting.

  • Agility AI. (2024) - Mastering Prompt Engineering. Episode 1 [summary of several standard industry frameworks]
    • Instructions – clearly and directly tell the AI what you want it to do (be direct and to the point)
    • Information – provide context and the data the AI needs to fulfill your instructions
    • Examples – show the AI what you are looking for in terms of style, form, format
    • Output format – how should results be shown

  • Lo, L.S. (2023). The CLEAR path: A framework for enhancing information literacy through prompt engineering, The Journal of Academic Librarianship, 49 (4). https://doi.org/10.1080/10875301.2023.2227621
    • CLEAR: Concise, Logical, Explicit, Adaptive, and Reflective

Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P.,

Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-

Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D.

M., Wu, J., Winter, C., … Amodei, D. (2020). Language models are few-

shot learners. Advances in Neural Information Processing Systems, 33,

1877–1901. https://arxiv.org/abs/2005.14165 (**not entirely attributed to

one set of authors)

30 of 31

Questions?

31 of 31

AI Prompting

For Faculty and Staff at USF

Prof. & Chair Allison Thorson, athorson@usfca.edu

April 7, 2026