Applying Advanced Generative AI Prompting Techniques

Emily Dux Speltz and Abram Anders
Iowa State University

This assignment is designed to enhance students' proficiency in advanced AI prompting techniques. The primary goal is to transform students who are already familiar with basic AI tool usage into advanced users capable of employing complex prompts effectively. Overall, we have found the activity to be successful in encouraging students to experiment with various advanced prompting techniques in a risk-free setting, allowing them to identify the prompting methods that resonated most with them. 

Learning Goals

  • Develop Advanced AI Prompting Skills: Students will develop a deep understanding and proficiency in using advanced AI prompting techniques. This includes being able to distinguish between different techniques, understanding their applications, and effectively employing them in varied contexts.
  • Apply AI in Real-World Scenarios: Students will gain experience in applying AI tools and techniques in practical, real-world-like scenarios—in our case, we chose marketing and business contexts, but this could be modified for other disciplines and applications. This goal focuses on translating theoretical knowledge into practical skills.
  • Engage in Creative Problem-Solving and Innovation: Students will think creatively and innovatively in using AI for various tasks. This includes personalizing prompts to suit specific scenarios, and devising unique and effective solutions to given tasks, thereby enhancing their creative problem-solving skills.

Original Assignment Context

This teaching activity took place in our "AI and Writing" course, an experimental elective offered during the fall semester by the Department of English at our university. The course was uniquely positioned at the intersection of technology and humanities, exploring the role of Artificial Intelligence in writing. It attracted a diverse group of undergraduate and graduate students from various disciplines, all looking to deepen their understanding and abilities to use AI for writing tasks. 

We assigned this activity during the ninth week of the semester, so students had already accumulated a good amount of experience using AI tools, having engaged with some prompt scaffolding and several basic techniques alongside strong prompting strategies. However, they hadn't yet explored these advanced techniques, making this the ideal time to introduce more complex and nuanced aspects of AI interaction in the writing process.

Materials Needed: Internet access and access to a generative AI tool such as ChatGPT, Claude, Bard, or Bing AI (Copilot)

Time Frame: One 50- to 80-minute class period

Overview

We introduced an activity designed to enhance students' proficiency in advanced AI prompting techniques. The primary goal was to transform students who were already familiar with basic AI tool usage into advanced users capable of employing complex prompts effectively. This activity was used in a university “AI and Writing” class setting in the Fall 2023 semester.

The activity began with a brief lecture that introduced 12 advanced prompting techniques. Each technique was accompanied by a sample prompt, offering tangible examples for application. We were interested in a relatable task that would allow students to translate theoretical knowledge of prompting techniques into practical skills, including multimedia outputs. In our case, we chose a task that would model marketing and business contexts, but this could be replaced with tasks and applications from other disciplinary contexts. Students were instructed to utilize advanced prompting strategies for four distinct marketing tasks: social media promotion, YouTube video creation, influencer outreach, and customer email communication. This hands-on approach required them to report the AI tools used, the prompting techniques attempted, and the AI-generated outputs.

The results were impressive. Students demonstrated a remarkable ability to apply advanced techniques creatively and effectively. For instance, one student, using the “Prompt the Model to Debate or Consider Multiple Perspectives” technique, generated comprehensive video ideas for a themed restaurant, while another employed the “Flipped Interaction Pattern” to script a promotional video for a car detailing business. Furthermore, in our observations, each of the recommended generative AI tools provided effective explanations of the prompting techniques which aligned with our lecture and Brownell (2023). This demonstrates the potential of AI as a teaching aid, particularly in enhancing students' abilities to interact with and utilize AI systems more effectively.

Overall, the activity proved successful in encouraging students to experiment with various advanced prompting techniques in a risk-free setting, allowing them to identify the prompting methods that resonated most with them. In subsequent course activities, we observed students making independent use of these techniques and adapting them to their unique needs. In final course reflections, students highlighted three advanced prompting techniques as particularly useful and versatile: iterative refinement, the flipped interaction pattern, and the question refinement pattern. In future iterations of this activity, we plan to incorporate reflective sessions in which students compare and evaluate the effectiveness of different techniques, further enriching the learning experience. This approach has not only broadened their technical skills but also fostered a deeper understanding of AI's potential in creative and business applications.

Assignment: Applying Advanced Prompting Techniques

Getting Started

Use a generative AI tool such as ChatGPT, Claude, Bard, or Bing AI (Copilot) to learn about the following advanced prompting techniques:

  1. Use Step-by-Step or Iterative Queries
  2. Ask the Model to Think Aloud
  3. Prompt the Model to Debate or Consider Multiple Perspectives
  4. Self-Correction Request
  5. Iterative Refinement
  6. Analogies and Hypotheticals
  7. Meta-Questions
  8. Feedback Loop
  9. The Flipped Interaction Pattern
  10. The Question Refinement Pattern
  11. The Cognitive Verifier Pattern
  12. The Reflection Pattern

Try asking for explanations and examples for each technique. Then, compare the results from two different tools. For techniques 9–12, you can compare your results to Brownell (2023). 

Objective 

  • Consider that you are running a business in town. It could be for a small artsy boutique, a fitness training program, a new restaurant, or anything that sounds fun to you.
  • Use advanced prompting techniques with an AI tool to help you develop a business marketing plan quickly and efficiently.
  • Complete Tasks 1–4 below with your imaginary business in mind. 

Tasks

Task 1: How can you market your business on Twitter, Facebook, and Instagram? 

  1. Which AI tool(s) did you use for this task?

  2. Which advanced prompting technique or series of techniques did you use for this task? Describe and provide your prompt(s) here.

  3. What ideas did the AI tool suggest for your business?

Task 2: How can you create a YouTube video to promote your business?

  1. Which AI tool(s) did you use for this task?
  2. Which advanced prompting technique or series of techniques did you use for this task? Describe and provide your prompt(s) here.
  3. What ideas did the AI tool suggest for your business?

Task 3: How can you reach out to influencers to share your products/services with their followers? 

  1. Which AI tool(s) did you use for this task?

  2. Which advanced prompting technique or series of techniques did you use for this task? Describe and provide your prompt(s) here.

  3. What ideas did the AI tool suggest for your business?

Task 4: How can you send out emails to your loyal customers? 

  1. Which AI tool(s) did you use for this task?

  2. Which advanced prompting technique or series of techniques did you use for this task? Describe and provide your prompt(s) here.

  3. What ideas did the AI tool suggest for your business?

Acknowledgements

We acknowledge Briana Brownell for documenting Techniques 9–12: 

Brownell, B. (2023, July 28). 5 advanced prompts to get better answers from ChatGPT. Descript. https://www.descript.com/blog/article/5-advanced-prompts-to-get-better-answers-from-chatgpt

These references also served as inspiration and examples for some of the prompting techniques that we introduced to students:

ChatGPT prompt engineering, ‘Let’s Think Step by Step’, and other magic phrases (2023, Feb. 10). Plainenglish.Io/Blog/Chatgpt-Prompt-Engineering-Lets-Think-Step-by-Step-and-Other-Magic-Phrases, Plain English. plainenglish.io/blog/chatgpt-prompt-engineering-lets-think-step-by-step-and-other-magic-phrases. 

Cheung, R. (n.d.). 500+ Best ChatGPT Prompts. Notion. Retrieved August 10, 2023, from https://vaulted-polonium-23c.notion.site/500-Best-ChatGPT-Prompts-63ef8a04a63c476ba306e1ec9a9b91c0 

Gallo, C. (2023, Feb. 23). How to use ChatGPT’s analogy feature to turbocharge your presentations. Inc. www.inc.com/carmine-gallo/how-to-use-chatgpts-analogy-feature-to-turbocharge-your-presentations.html. 

We used ChatGPT 4.0 to generate examples of prompts and sample outputs for students in class.