How to Write Prompts for AI: A Student Guide

Time to Read: 10 minutes

Good AI prompts are specific. Tell the tool what you are trying to do, give it context, ask for a clear output, and then follow up when the answer is too vague.

For students and early professionals, the useful part is not sounding technical. It is knowing how to use AI for research, studying, job applications, interview prep, and career exploration without letting it think for you.

If you are using AI for career research, pair it with real-world information. Use it to prepare better questions before reading Explore Career Paths, working through Getting Hired, or talking to someone already doing the work.

Basics of Writing AI Prompts

Crafting effective prompts for AI is both an art and a science. You’ll need to learn how to communicate your needs clearly and concisely to get the best possible output from AI tools like ChatGPT. Here’s how to approach writing AI prompts:

Understanding Your AI Companion

Before diving into prompt writing, recognize that AI models process information and generate responses based on patterns learned from vast datasets. Your goal is to guide the AI with clear, direct prompts.

Approaching AI interactions conversationally, rather than as mere transactions, transforms the dynamic from simply asking and receiving to engaging in a meaningful dialogue. This conversational approach mirrors human learning interactions—rich with questions, clarifications, and explorations—that naturally foster a deeper understanding. Even ChatGPT thinks so!

PRO TIP: Many users make the mistake of treating AI purely as a Q&A machine. It’s far more effective to approach your interactions in a conversational way, much like you would with another human. 

Clarity is Key

Start with clear, straightforward language. Ambiguity can lead to unexpected results, so specify exactly what you’re asking for. If you’re seeking an explanation on a topic, instead of saying “Tell me about quantum physics,” try “Explain the basic principles of quantum physics in simple terms.”

Be Specific

The more specific your prompt, the more tailored the AI’s response will be. For research-related prompts, instead of a broad request like “How do I write a research paper?”, specify the subject, “What are the steps to write a research paper on renewable energy technologies?”

Tips for Improvement

Letting AI “Think” in Prompt Engineering

When we talk about letting AI like ChatGPT “think out loud,” it’s not about AI thinking in the human sense but rather about making its processing steps explicit in its response.

An easy way to let AI “think” out loud is to get it to “reflect” on its answers. In doing so, you encourage a level of self-assessment that can reveal the reasoning behind its conclusions. This practice is proven to lead to better quality responses.

Getting ChatGPT to “reflect” on its answers lead to a 14.58% increase in the accuracy of its responses when tasked with answering complex reasoning questions.

Geng et al. 2023

For instance, after receiving an initial explanation, prompting the AI with, “Reflect on why this solution is optimal,” or “Can you explain how you got to this conclusion and what makes it better than other alternatives?” can lead to richer, more nuanced insights. This step is especially important for academic prompts where understanding the ‘how’ and ‘why’ is as important as the answer itself, fostering a deeper engagement with the material.

Iterate and Refine

Don’t expect perfection on the first try. Use the AI’s responses to refine your prompts, adding more detail or adjusting the focus as needed. View each interaction as part of an ongoing conversation. If an AI’s response opens new points of interest or raises further questions, don’t hesitate to inquire about them in subsequent prompts. This iterative process mimics the natural way we deepen our understanding through discussion.

Use Feedback Loops

After receiving a response, if something isn’t clear or you think the AI can provide more information, it is good practice to ask follow-up questions. For instance, “Based on your previous response, can you dive deeper into the implications of rising sea levels on coastal cities?” Feedback for AI is important because it helps in refining the accuracy and relevance of its responses, ensuring they meet the users’ needs more effectively.

PRO TIP: You can ask ChatGPT for feedback on your prompts by explicitly requesting insights on how to improve them for clearer, more effective communication, or how to better structure your questions to receive the most relevant and accurate responses.

Explore Different Phrasings

Sometimes, rephrasing a question can lead to better responses. If the initial prompt doesn’t yield the desired result, try approaching the question from a different angle. It’s similar to how you would ask for clarification when in conversation with a human. 

Prompt Engineering Framework: ACTOR

The ACTOR framework is a structured approach for how to write prompts for AI, designed by GradSimple for college students. It ensures clarity, context, and purpose in your interactions. Each component—Assume, Context, Task, Objective, Result—plays a critical role in crafting effective prompts that yield relevant and actionable AI responses. Let’s delve into what each alphabet represents and its importance, using the context of a college student seeking help from an AI math tutor.

Assume (A) – Define the Role

Importance: Specifying a role for the AI sets the tone and direction of the response. It helps the AI adopt a specific perspective or expertise, aligning its responses with the expected knowledge base or skill set.

Context (C) – Provide Background

Importance: Context enriches the prompt with relevant background information, enabling the AI to tailor its responses more accurately to the user’s situation. It includes details about the user’s current understanding, specific challenges, and any relevant circumstances.

Task (T) – Specify the Action

Importance: Defining the task directs the AI on what exactly needs to be done. It transforms the context and role into actionable items, specifying the type of assistance or information required.

Objective (O) – State the Purpose

Importance: Articulating the objective clarifies the purpose behind the prompt, guiding the AI to understand the ultimate goal of the task. It ensures the response is not just accurate but also aligned with the user’s needs.

Result (R) – Describe the Desired Outcome

Importance: Detailing the desired outcome in terms of format, structure, or content ensures that the AI’s response meets the user’s expectations. It adds a layer of specificity regarding how the information should be presented.

Challenges and Solutions in Prompt Engineering

Challenge: Difficulty in Formulating Effective Prompts

Crafting prompts that yield the desired AI response often involves trial and error. New users may find it challenging to balance specificity with clarity, leading to outputs that don’t fully meet their needs.

Solution: Embrace the iterative nature of prompt engineering. Start with any prompt and refine it based on AI’s responses. For instance, if an initial prompt about “the history of the internet” generates too broad an overview, narrow it down with, “Focus on the development of internet protocols in the 1980s.” This process of adjustment helps hone your prompt-crafting skills over time.

Challenge: Vague or Broad Responses

Sometimes, AI might provide responses that are too general or broad, lacking the depth or specificity needed for academic work.

Solution: Enhance prompt specificity by including detailed questions and context. For example, instead of asking, “What are renewable energy sources?” specify, “List renewable energy sources suitable for urban environments and discuss their efficiency and implementation challenges.”

Challenge: AI “Hallucinations” or Misinformation

AI models, including ChatGPT, can sometimes “hallucinate” information—producing plausible but incorrect or misleading data.

AI models generate responses based on statistical likelihoods—patterns learned from vast datasets, rather than accessing real-time, verified facts.

Solution: Combat this by cross-referencing AI responses with credible sources. For research-intensive tasks, follow up with, “Provide sources or evidence to support your claims on the impact of climate change on polar bear populations.”

Challenge: Misalignment with Academic Standards

AI-generated answers might not always adhere to the specific standards or depth required for college assignments.

Solution: Explicitly incorporate academic guidelines into your prompts. For a literature review, instead of a vague request, specify, “Summarize key findings on behavioral economics from articles published in the last five years, adhering to Chicago style formatting.” This directs the AI to produce content that aligns with academic conventions.

Challenge: AI Getting “Distracted” or Derailed

Sometimes, certain elements within a prompt can lead AI down an unintended path, producing responses that veer off-topic or focus on less relevant aspects.

Solution: Be mindful of the specificity and focus of your prompts. If an initial request for information about “sustainable practices in urban planning” leads to a generalized discussion on sustainability, refine the prompt to “Identify specific sustainable practices in urban planning that reduce carbon footprint, focusing on case studies from European cities.” This directs the AI to remain focused on the core topic and avoid tangential content.

How to Write Prompts for AI: Concluding Thoughts

The ACTOR framework is a useful starting point because it forces you to give AI a role, context, task, objective, and expected result. That is usually enough to turn a vague answer into something you can use.

For career decisions, use AI to organize your thinking, draft questions, compare options, and prepare outreach. Then check the answer against real people, real jobs, and real hiring expectations. AI can help you think faster, but it should not be the only source you trust.

If you are using AI to research a path with real stakes, GradSimple Insiders can help you compare the answer against what people inside those careers are actually seeing.