Generative AI: Prompting

A prompt is a set of instructions for a language model to perform a task. Tasks can take many forms, including text generation, image creation, translation, summarizing, writing, etc. A prompt might consist of a few words, a single sentence, or extended paragraphs.  The way prompt has a huge impact on the quality of the results you receive.

When crafting a prompt:

• Use plain English and specify what you want the tool to do.
• Providing background information, context, or examples can be helpful.
• Divide a complex concept into more minor prompts.
• Ask ChatGPT to ACT AS AN EXPERT to explain a complex concept step by step.
• If you want to change to an unrelated topic, click NEW CHAT. But if you want to keep the previous conversations together in one chat, you could use a prompt like "Ignore the previous conversations. Let me ask you a new question."
• Choose an output format. In addition to paragraphs it can give you a table, a bulleted list, ascii art, multiple choice quiz questions, emojis, computer code, and more.

Crafting a well-formulated and specific prompt to interact with Generative AI can improve your results.  Guidelines from the CLEAR Framework for Prompt Engineering provide these things to include and consider when crafting a prompt:

• Concise-brevity and clarity in prompts
• Logical - structured and coherent
• Explicit - clear output specifications
• Adaptive - flexibility and customization
• Reflective - continuous evaluation and improvement

Source: Lo, Leo S. “The CLEAR Path: A Framework for Enhancing Information Literacy through Prompt Engineering.” Journal of Academic Librarianship, vol. 49, no. 4, July 2023, p. N.PAG. EBSCOhost, https://doi.org/10.1016/j.acalib.2023.102720.

Below are examples of AI prompts an instructor might use in their teaching.  The effectiveness of prompts may vary based on the specific AI model, context, and platform you're using. Flexibility, experimentation, and ongoing refinement are key aspects of mastering the art of prompt writing. Click on the example to view the prompt and output in ChatGPT.

• Example 1: Create a syllabus Create a syllabus for a 15-week introductory, undergraduate-level sociology course that meets once a week for 90 minutes. The syllabus should outline the course objectives, topics, grading policies, guidelines for using generative AI, and expectations for student behavior and participation. The required textbook for the course is Anthony Giddens et al. Introduction to Sociology, 12th edition. For each week, suggest lecture topics, readings, and assignments based on the chapters of the textbook.

• Example 2: Design a lesson plan Design a ninety-minute lesson plan for an undergraduate American history course on the 1886 Haymarket Square Riot. The plan should include activities that help students understand the factors leading to the event, its consequences, and its relevance to the contemporary labor movement. It should also incorporate primary sources, classroom discussions, and multimedia resources designed to engage students in critical thinking and analysis.

• Example 3: Suggest discussion topics Generate five discussion prompts about Howard Zinn's A People's History of the United States for a graduate-level U.S. history course. Generate one prompt for each of the following topics discussed in the book: colonization, slavery, labor movement, black civil rights, and women's liberation. Each prompt should ask students to engage with at least one primary source about that topic.

• Example 4: Create a quiz Create a 10-question multiple choice quiz for an undergraduate introductory physics course based on the following learning objective: Explain the properties of waves, including wavelength, frequency, amplitude, and velocity. Indicate the correct answer for each question.