By the end of this unit, you will be able to apply key principles to craft effective prompts for Generative AI (GenAI) tools, like CUNY Copilot, ChatGPT, Claude, Gemini, etc. to support your academic, research, or administrative tasks responsibly.

Components of a Good Prompt

Effectively interacting with Generative AI (GenAI) tools requires more than simply typing a question into a chat box. Unlike search engines, GenAI responds based on your prompt’s structure, context, and clarity, not just keywords. Prompting is not “googling”; it is a skill that involves intentionally guiding the AI to produce thoughtful, relevant, and usable results. This section outlines the essential components of a good prompt—a practical set of strategies you can apply to improve the quality, reliability, and academic appropriateness of GenAI outputs. You can mix and match these building blocks to generate more accurate, relevant, and context-aware outputs.

Example

Act as an experienced university-level instructional designer specializing in creating engaging online discussions.

I am teaching an undergraduate online asynchronous course titled 'Digital Ethics in the 21st Century.' The current module is 'Algorithmic Bias and Social Justice.' Students have already read chapters on the definition of algorithms, sources of bias in AI, and case studies of discriminatory algorithms in hiring and criminal justice. They have a foundational understanding but need to engage more deeply with solutions and personal reflections.

Your task is to generate one thought-provoking discussion forum prompt. This prompt should encourage students to critically analyze potential solutions to algorithmic bias and reflect on their own roles and responsibilities in addressing these issues.
Please structure your output as follows:
-Discussion Prompt Title: (A concise and engaging title)
-Prompt Instructions: (The full text of the prompt for students, including guiding questions)
-Learning Outcomes Addressed: (List 2-3 specific learning outcomes this prompt would help achieve, related to critical thinking and ethical reasoning)
-Suggested Grading Criteria (Brief): (Provide 2-3 key elements a TA might look for when assessing student responses, e.g., depth of analysis, connection to course materials, constructive engagement with peers). Present these criteria as bullet points.
The prompt instructions should be between 150-200 words.

It must explicitly ask students to propose at least one concrete solution or mitigation strategy for algorithmic bias.It should also prompt students to consider the ethical responsibilities of different stakeholders (e.g., developers, policymakers, users).
Avoid overly technical jargon; the language should be accessible to undergraduates from diverse disciplinary backgrounds.

Do not suggest solutions that have already been extensively covered in the hypothetical readings mentioned above. Focus on prompting novel thinking or deeper application.

Prompting AI Agents

In addition to the core components of a good prompt, specifying a structured output format can significantly improve the usability of AI-generated responses, especially when writing instructions for AI agents. These agents are specialized AI tools designed to assist with specific tasks, such as drafting content, summarizing information, or generating step-by-step guides.

One particularly effective format to request is Markdown. Markdown is a lightweight plain-text formatting language that uses simple symbols to create structure, such as headers, bullet points, bold text, and links.

# stands for top-level header

## is a second-level header

** bold text

– bullet polint

Including a request for Markdown in your prompt helps the agent deliver well-formatted, ready-to-use output.