Using Generative AI (such as ChatGPT) for Research & Writing

When using AI for your research, it is only as useful as the questions you ask and how you interact with it. To generate accurate, detailed, and relevant responses, be clear and direct with your initial question, evaluate AI’s response, and provide clarifications, context, and follow-up prompts.

1. Ask: Pose your initial question or request.
2. Evaluate: Read the response. What’s useful? What’s missing?
3. Revise and Reask: Adjust your prompt. Add context, clarify the task, or set a new constraint.

When creating your initial ask, and when creating any follow-up prompts based on AI’s response, consider including COSTAR information:

• Context (C): Provide background information (ex: “My research topic is on …”)
• Objective (O): Clearly define the task (ex: “Generate a list of keywords...”)
• Style (S): Specify the writing style (ex: “…in a persuasive style…”)
• Tone (T): Set the mood (ex: “...using a professional tone…”)
• Audience (A): Identify your audience (ex: “The audience is someone new to academic research.”)
• Response (R): Identify the format you want (ex: “Generate a table of pros and cons…”)

To see this in action, review the BASIC, BETTER, and BEST examples below:

Basic Prompt Example: “Tell me about school lunch programs in the US.”

Better Prompt Example: “Provide me with a summary of the pros and cons of US public school lunch programs.”

Best Prompt Example: “My research topic is on the school lunch program in the United States. Using a persuasive style, generate a table that provides the pros and cons of public-school lunch programs. Your audience is someone who has no prior knowledge of the school lunch program in the United States.”

Pro Tips for Prompting AI:

Be clear and specific. Think critically about what you want and how you want AI to think about the topic and provide a response. Type in more than just key terms.

For example, “Explain three strategies leaders use to build trust in remote teams, with one real-world example for each.”

Provide AI with the situation or audience. 

For example, if you want information on teaching and learning, tell AI the situation and the audience you are thinking of: “What strategies can online teachers use to better engage middle school students?” 

Engage in a back-and-forth conversation with AI, providing clarifications, more context, and follow-up questions. You can ask for examples, or ask AI to clarify a point, or ask situational questions.

For example, if you ask AI about emotional intelligence, you can then ask for examples within a situation: “Can you give me an example of emotional intelligence in a workplace conflict?”

Break up questions to get a more detailed response. Instead of asking “What is vaping and why is it so dangerous?” perform two separate searches.

For example, “What is the history of vaping?” first, and then “How does vaping impact health?” 

 Help AI avoid misinterpretations when searching topics that might be ambiguous. If you want information about Apple, specify that you want information about the company and not the fruit.

For example, “…Apple the company and not the fruit….”

Include how you want the response to be provided. Do you want a summary or a detailed explanation? Do you want a bulleted list? Do you want the response to be prioritized in some order?

For example, “Give me a detailed bulleted list of the 3 most misunderstood things about space travel.”

Consider stating what you want it to avoid.

• For example, “Only provide information that was published between 2022-2026.”
• For example, “Do not re-write my paper. Instead, provide suggestions on areas where I have and have not addressed the assignment prompt.”

Provide AI with the role you want it to take.

• For example, “You are a career coach. Write a 2-paragraph explanation for a first-year college student about why time management matters.”
• For example, “Answer this as if you were a hiring manager deciding between two candidates.”

Use AI to discover gaps in your thinking, challenge your thinking or assumptions, offer alternative arguments, better organize your ideas or your paper, and **discover sources. See the "Get Immediate Writing Feedback or Academic Tutoring" section. 

• For example, “What might I be overlooking here?”
• For example, “Explain step by step how you arrived at this answer.”
• For example, “Play devil’s advocate. What are some weaknesses in this argument?”
• For example, “Here’s my academic paper. Suggest ways I could better organize these ideas.”
• For example, “What credible websites or organizations provide up-to-date statistics on global warming?"

**AI does not have access to all scholarly articles. Articles that require a subscription and are therefore not publicly available on the internet will not be included in the information generated by AI.  The university library holds subscriptions to these articles so you will still need to use the library for current, unbiased, and scholarly articles.  

**All sources and information generated by AI must be verified by you. There are times when generative AI provides a “source” that does not exist at all. You can only cite a source when you have located the information within that source yourself. And you should never cite AI as a research source.

A Note About Privacy

Avoid sharing information with any generative AI tool unless you're comfortable with the possibility that it could be incorporated into the model's understanding and used in subsequent interactions with other users.

Self-Check: Keeping a Human in the Loop

AI can generate ideas, structure information, and spark new directions but you are still the thinker and decision-maker. Before you use any output from AI, pause and consider:

• Usefulness: Did this output help me understand the topic better—or just give me something to turn in?
• Currency: Is this information up to date?
• Relevance: Does the information address what I need?
• Accuracy: How will I verify and cite information and sources from AI? See the "Verifying Information & Sources from AI" and "Disclosing & Citing AI" sections. 
• Authenticity: Do I agree with the reasoning here, or do I need to challenge it?
• Originality: Where can I add my own interpretation, perspective, and conclusions I’ve reached? See Academic Voice and the "Include Your Originality When Collaborating with AI" section.

Using AI responsibly means supporting your own critical thinking, not outsourcing it. The best results happen when you stay in charge, letting AI expand your perspective while you provide the judgment and integrity.

Verifying and vetting information is the most important step in effectively using generative AI.  Sometimes, AI produces information that sounds convincing but is simply wrong. This is called a hallucination. Examples might include:

• Quoting articles or studies that don’t exist.
• Confidently giving incorrect statistics.
• Misrepresenting a theory or author’s position.

A human must stay in the loop. As a student, it is your role and responsibility to check every AI response before using it for a discussion or assignment. As a career professional, the same is true for any work you submit for your job. As a civic person, the same is true for any communications you share with others.  

Is the information accurate? Always go to the source itself that generative AI is using to ensure it is an existing source, one that can be used for academic purposes, AND that the source includes the information you wish to use. If AI does not cite its source, you must locate a source that can verify the information. And you’ll then use and cite that as your paper’s source.

Does AI’s source actually exist? There are times when generative AI provides a reference to a “source” that does not exist at all. This is another reason to always go to any source that generative AI provides. You can only cite a source when you have located the information within that source yourself. And you should never cite AI as a research source. See information below and Verifying a Source. 

Is it credible, unbiased, and up to date? Verifying information isn’t just about ensuring accurate information. Generative AI may provide information that is accurate, but it may be providing this information from a source that is not highly credible or one that is biased or too old and no longer relevant. Only highly credible sources, such as scholarly sources, can be used for any academic research and writing you do.

If you include incorrect information provided by AI in an assignment, you will be responsible for it. Including non-existing and unverified sources in a references list is known as falsification. Providing inaccurate or misleading information in your paper is known as misrepresentation. Both are academic integrity violations that can easily happen when attempting to gather information from generative AI without also going directly to the/a source to verify it.

How To Verify Information & Sources

Track information back to the original source. Go directly to any source provided by AI to confirm it exists and is appropriate. For example, use the link icons in Gemini or the numbered footnotes in Copilot to see the source, then go to that source. If it is a credible source, utilize that author’s information instead of the AI-generated content. You would cite that source.

Check the dates of sources when verifying information from AI to ensure it is up to date. This is especially important for topics that are rapidly evolving.

Check with an authoritative source if AI a.) cites a source that is outdated, b.) cites a source that is not an academic or government source or c.) simply does not cite a source. Authoritative sources include academic journals or government and university publications. To determine authoritative and credible sources, use the Library’s Hierarchy of Sources guide.

Consult multiple sources to verify information rather than a single source.

Use fact-checking tools. Websites like Snopes, FactCheck.org, and PolitiFact can help you identify misinformation or inaccuracies.

Remember, AI can be a good source for initial research on a topic, but it cannot be the single research method you use. Do not cite AI as your source of information—only cite a credible source and only after you have verified it.

Self-Check: Keeping a Human in the Loop

AI can generate ideas, structure information, and spark new directions but you are still the thinker and decision-maker. Before you use any of AI's outputs, pause and consider:

• Usefulness: Did AI's output help me understand the topic better—or just give me something to turn in?
• Currency: Is the information up to date?
• Relevance: Does the information address what I need?
• Accuracy: Did I verify the sources AI used as highly credible sources that did include the information that AI generated? See the "Disclosing & Citing AI" section.
• Authenticity: Do I agree with the reasoning here, or do I need to challenge it?
• Originality: Where can I add my own interpretation, perspective, and conclusions I’ve reached? See Academic Voice and the "Include Your Originality When Collaborating with AI" section. 

Using AI responsibly means supporting your own critical thinking, not outsourcing it. The best results happen when you stay in charge, letting AI expand your perspective while you provide the judgment and integrity.

Creating your own original content is mandatory.

AI responses are generated by a bot, synthesized from information found readily online. Even if using the information from it, you will need to do significant editing and paraphrasing and you will need to cite the original source of any research information that AI provides. You should not cite AI as your source. See the "Disclosing and Citing AI" section. 

Any work you submit must be written by you. Using AI to generate text and then submitting it as your own work is considered plagiarism. AI can be a collaborator of ideas, but it is not meant to communicate for you. Your instructor should see your thinking—not AI’s output.

Spotting Over-Reliance (and Possible Academic Integrity Violations)

It’s easy to slip from “partnering with AI” into “outsourcing your thinking.” Signs you may be leaning too much on AI include:

• Turning in content you didn’t really read or understand.
• Copying text without paraphrasing or editing and adding your own insights or analysis.
• Feeling like you can’t complete an assignment without AI’s help.
• Using AI to replace (instead of support) skills you’re expected to practice such as writing, problem-solving, or interpreting data/findings.

Pro Tips to Maintain Your Original Content and Voice:

Do not copy/paste AI text directly into your paper (this also includes text generated by any paraphrasing software). This is an academic integrity violation because you did not write the text.

Keep your voice. AI often writes in a style that probably doesn’t sound like you. Before using its output, revise for tone, clarity, audience, and authenticity. Add your own transitions, word choices, and personal reflections so the final product reflects your voice. A good test: if someone who knows you read it, would they recognize it as yours?

Expand your own perspective. Think of AI as a brainstorming partner that can generate ideas you might not have considered. But don’t stop there. You bring the context, emotional intelligence, and understanding of your audience. Decide what generated content is relevant for your uses. Use the relevant content as points of discussion but then elaborate on those points of discussion with your own writing. Add your own insights and context to make it your own.

• For example, if AI suggests five strategies, maybe you take two and connect them to your course readings.
• For example, AI might generate a list of pros and cons, but only you can decide which ones matter most for the assignment or which align with your values.

Do not directly quote AI as if it is an author. If AI cites a source in their response to your question, locate that source to first verify it exists and paraphrase or quote that author directly if it is a reliable source. If no source is provided by AI, locate a reliable source to verify the information and then use and cite that source.

Remember, AI can give you some excellent ideas to consider, but it is your responsibility to create the final communication or writing to avoid plagiarism and preserve your integrity.

Self-Check: Keeping a Human in the Loop

AI can generate ideas, structure information, and spark new directions but you are still the thinker and decision-maker. Before you use any of AI's output, pause and consider:

• Usefulness: Did this output help me understand the topic better—or just give me something to turn in?
• Accuracy: How will I verify and cite information and sources from AI? See the "Verifying Information & Sources from AI" and "Disclosing & Citing AI" sections. 
• Originality: Where can I add my own interpretation, perspective, and conclusions I’ve reached? See Academic Voice. 
• Ethics: Does this align with the standards of my course and my own ethics?

Using AI responsibly means supporting your own critical thinking, not outsourcing it. The best results happen when you stay in charge, letting AI expand your perspective while you provide the judgment and integrity.

You may need to cite AI or include a disclosure statement about how you used it in your assignment.

AI is a source of information and to include any content from it without disclosing it is a form of dishonesty. In the context of academic writing, it is a form of academic dishonesty or plagiarism.

Pro Tips for When to Disclose or Cite Generative AI:

Cite AI’s output directly only if you are using it for a paper specifically on AI (such as “the limitations of generative AI”). You can use and cite some of AI’s responses as evidence on a paper about generative AI. See the "Citing Generative AI (Such as ChatGPT)" section within APA: Citing Within Your Paper.

Do not cite AI as if it is an author when using it as a source of information. If AI provides a source in their response to your question, locate that source to verify it exists and cite that author directly if it is a reliable source. If no source is provided by AI, locate a reliable source to verify the information and then use and cite that source. See the "Verifying Information & Sources from AI" section.

You should include a disclosure statement if you use AI during the pre-development process to help you narrow your topic, identify ideas, develop an outline, or for revising and editing your work. See section below. You do not need to include the use of grammar-checking tools, citation generators, or plagiarism detectors within a disclosure statement. 

Where & how to include a disclosure statement:

Include a disclosure statement at the end of your paper, PowerPoint, etc. before your list of references. If you are a doctoral student writing your dissertation or applied doctoral project, the disclosure statement should be included within your Methodology chapter or section. Do not include AI within your list of references. State how, when, and to what extent AI was used in your research or writing process. For example: "I used ChatGPT to narrow my topic and to generate some key topics for this paper." 

Remember, using AI to write any part of your paper for you is an example of plagiarism and academic dishonesty even if you cite it or disclose it. Do not use AI in this way. See the "Include Your Originality When Collaborating with AI" section. 

Generative AI can be an alternative option for academic support or immediate writing feedback. AI can be a patient, encouraging, and on-demand tutor or teaching assistant, but you’ll have to prompt it to provide meaningful and helpful teaching advice.

Pro Tips for getting AI to act as an academic tutor or teaching assistant:

Provide a specific and directive prompt for what you want the tool to do. Tell it what its role is, what your role is, and how you want it to interact with you. For example:

Ask AI to quiz you on a topic. Provide a prompt telling it that you are a student and to ask you questions about a topic, letting you know when you get a question wrong or right. Tell it to provide instruction when you get it wrong and to make the questions progressively harder as you continue to get correct answers.

Get help with a written assignment. Copy and paste this example prompt into the generative AI tool to get specific writing feedback:

“You are an academic writing tutor whose role is to give direct and clear instruction and feedback on a student’s written work and to ask the student to explain your feedback and how they plan to apply it. Answer any questions the student has about your feedback. First, ask the student what type of writing they are working on and the topic of it. After the student has responded, ask them to provide the rubric or the assignment instructions if they can. After the response, ask them to copy and paste their written work for you to review. Once you have their written work, assess their writing and provide feedback based on the assignment instructions or rubric and all you know about the expectations for the type of writing. Provide a balanced overview of their work, pointing out strengths and areas for improvement. End by asking if the student has any questions about your feedback and how they plan to use it. If the student has questions, ask them what they think is the answer before you provide the answer.”

Get helpful academic instruction. Copy and paste this example prompt into the generative AI tool:

“You are an encouraging academic tutor who helps students to understand concepts by explaining and asking the student questions. Ask the student what they want to learn about. When explaining concepts, provide examples and analogies. Ask leading questions to have the student engage with you about the concepts and their understanding of it. Ask the student to explain their thinking and if they get the concept wrong, provide a hint. When the student shows understanding of the concept, praise them and let them know that they understand and can beginning working on their own.”

Example prompts in this section were adapted from E. Mollick & L. Mollick. (2023). AI as feedback generator & Feedback as personal tutor. Student Use Cases for AI: An Inspiring Minds Series. https://hbsp.harvard.edu/inspiring-minds/student-use-cases-for-ai