爱污传媒

Skip to Main Content Skip to bottom Skip to Chat, Email, Text

How to Write an Effective Prompt for Generative AI Tools

High school teacher leading a blended learning class

By Mansureh Kebritchi, Ph.D.

Introduction

Developing effective prompts for generative AI tools is an essential skill that plays a crucial role in determining the quality and relevance of the generated output. Crafting compelling prompts requires a thoughtful approach that takes into account clarity, specificity, and context. You may use the general tips and some of the prompt engineering techniques discussed below to create more effective prompts.

General Tips for Effective Prompting听

1.听听听听听 Be Specific and Clear

Use precise language and avoid ambiguity. Clearly state what you want the AI to generate, whether it is a text, an idea, or a solution. Specific prompts yield more focused results.

  • Avoid Ambiguity: Being specific reduces the chance of misinterpretation. For example, instead of saying "Write a story," try saying "Write a short story about a dog who discovers a new passage to his home."
  • Define the Format: Specify the desired output formats such as code, script, email, list, table, story, or poem, as needed.
  • Specify the Length: Give an indication of the desired length, for example, specify, 听"Write a 250-word essay" or "Write a 10-line abstract".

2.听听听听听 Provide Context and Background

Offer background information or context to help the AI understand your request better. Including relevant details will guide the AI in producing more accurate and nuanced responses.

  • Set the Scene: If the AI tool requires understanding a specific context, provide it. For example, specify "You are a marketing consultant. Create a catchy phrase for a new laptop brand."
  • Define the Audience: Identify the intended audience for your writing. For example, prompting, "Create a product description for a new smartphone targeted at teenagers" will generate different results than one aimed at educators.
  • Add Constraints: Add limitations to the AI's output, for example, state, "Write a poem about a beach, but do not use the words 'red' or 'Sand.'"

3.听听听听听 Use Keywords and Phrases

Using keywords and phrases can enhance your communication with AI tools, resulting in more relevant outputs.

  • Identify Key Concepts: Include relevant keywords that are essential to the topic.
  • Use Strong Verbs: Use strong action verbs to guide the AI's focus. Consider words like "analyze," "summarize," "create," "compare," and "contrast," depending on your context.
  • Specify the Tone/Style: To better describe your intended tone use words such as "formal," "informal," "humorous," "serious," "persuasive," or "technical".

4.听 听 听 Iterate and Revise

Experiment with different versions of your prompts. If the initial results aren鈥檛 satisfactory, tweak your wording or provide additional context until you achieve the desired outcome.

  • Start Simple: Begin with a basic prompt and gradually add more detail.
  • Experiment: Try different phrasings and keywords to see what works best.
  • Analyze the Output: Review the results carefully and adjust your prompt as needed. If the output is incorrect, try rephrasing or adding more context.
  • Use Feedback Loops: Many AI tools allow you to provide feedback (thumbs up/down) on the generated content. This helps the AI tool learn your preferences.

Adopt Prompt Engineering Techniques

Prompt engineering provides techniques to enhance the output from your AI. Some of these techniques include:

  • Zero-Shot Prompting: Zero-Shot promoting asks AI tools to perform a task without providing examples within the prompt. It is useful for situations where examples are not available/unnecessary such as translation, summarization, or analysis. 听
  • Few-Shot prompting (Give Examples): Few-shot prompting is a technique that involves including examples in the prompt to help AI tools learn more effectively. This approach assists the AI in understanding what you are looking for. Few-shot prompting is especially beneficial for complex tasks where zero-shot prompting might not generate satisfactory results.
  • Chain Prompts: Break down complex tasks into smaller, sequential prompts. For example, first ask the AI to generate ideas, then ask it to elaborate on the best ideas.
  • Role Playing: 听The Role-Playing technique instructs the AI to take on a specific role to complete the assigned tasks more effectively. For example, you might prompt the AI by saying, "You are a travel agent. Plan a 3-day trip to New York for a family of five."

Conclusion

By following the above tips, you can significantly enhance the quality and relevance of the output generated by AI tools, specifically keep the following in mind as you develop your prompt:

  • Specificity is crucial. 听The more details you provide, the better the results you will receive.
  • Iteration is essential. Do not expect to achieve the perfect output on your first attempt.
  • Prompt engineering is a valuable skill. Practicing and experimenting will improve your abilities in this area.

Note that prompt engineering is an ongoing learning process, so don鈥檛 hesitate to experiment and try new approaches.

Resources

(2023). Prompt Engineering & AI Institute.

For questions and additional guidance join听Al Research Group

Mansureh Kebritchi, Ph.D.

ABOUT THE AUTHOR

Mansureh Kebritchi, Ph.D., is an accomplished educational researcher with over 5000 peer-reviewed citations. She is an expert in instructional design, project management, and dissertation/research mentoring. Dr. Kebritchi is the founder and chair of the听Center for Educational and Instructional Technology Research听at the College of Doctoral Studies, 爱污传媒. In addition, she is the founder and leader of听Dissertation to Publication听and听Research to Publication听Workshops,听Research Methodology Group, and听Al Research Group. She has served as the guest editor for the 爱污传媒 Scholar Special Editions on听Adult Learners in Online Higher Education听and听Reshaping Education through Innovation. She has a deep passion for educational research and extensive experience as a research methodologist and instructional designer. Dr. Kebritchi is dedicated to mentoring doctoral students and supporting faculty members in their pursuit of conducting and publishing research in the field of education.听Dr. Kebritchi is passionate about finding new and creative methods to enhance the quality of education in different settings, whether it be K-12, higher education, or corporate environments, both online and in-person.听