Polk Library

Purpose: Use AI carefully during early exploration so that it supports your thinking without taking over your topic, question, or direction.

Decision Point 1: AI Engagement

Guiding question: When and why should I use AI?

 

What research suggests: AI can effectively support bounded activities such as brainstorming, clarification, feedback, and language refinement. Selective use may reduce lower-level linguistic demands while allowing researchers to remain engaged in planning, reasoning, and reflection. In contrast, frequent or uncritical reliance can reduce originality, cognitive effort, and authorial control and may contribute to formulaic writing.

Implication for research practice: Researchers must determine whether AI is supporting their thinking or displacing the intellectual work necessary to develop their ideas.

Guidance for practice:

• Establish your purpose and preliminary direction before relying extensively on AI.
• Use AI selectively for clearly defined tasks, such as testing ideas, clarifying concepts, soliciting feedback, or refining language.
• Evaluate AI suggestions rather than accepting them automatically.

Decision Point 2: Problem Framing

Guiding question: How is AI shaping my understanding of the topic?

 

What research suggests: AI can help researchers explore unfamiliar concepts, identify perspectives, and generate possible directions. Its responses, however, may emphasize conventional interpretations, generic structures, or incomplete explanations. These suggestions can influence which aspects of a topic appear relevant or worth pursuing.

Implication for research practice: Researchers must recognize that AI-generated explanations and suggestions can affect how they define a problem, even when the output is used only for initial exploration.

Guidance for practice:

• Begin with your own questions and exploratory thinking.
• Treat AI-generated explanations as possible perspectives rather than complete accounts.
• Compare AI suggestions with scholarly sources, disciplinary concepts, and alternative interpretations.
• Check whether your framing reflects your inquiry or follows a generic structure suggested by the tool.

Decision Point 3: Relevance and Positioning

Guiding question: What information is worth keeping, and how does it fit?

 

What research suggests: Productive AI use depends on disciplinary knowledge, metacognitive awareness, and the ability to evaluate outputs for accuracy, relevance, and contextual fit. Fluent or topically related information may still be inaccurate, irrelevant, or poorly positioned within an argument.

Implication for research practice: Researchers must decide not only whether an output relates to the topic, but whether it contributes meaningfully to the research question, argument, and disciplinary context.

Guidance for practice:

• Evaluate AI-generated information against your research goals and disciplinary knowledge.
• Ask what role each idea would play in your argument and what evidence supports it.
• Reject material that merely repeats the topic, adds generic content, or appears relevant only because it is presented fluently and coherently.

Purpose: Use scholarly sources, library databases, and disciplinary standards to verify AI-generated information before relying on it.

Decision Point 4: Accuracy

Guiding question: Is this information correct?

What research suggests: AI systems can provide useful and correct information, but they can also produce factual errors, misleading interpretations, and unsupported claims. AI assistance may increase the amount of correct information incorporated into writing without necessarily preventing incorrect information from being included. Metacognitive awareness and active evaluation therefore remain essential.

Implication for research practice: Researchers remain responsible for verifying factual claims, no matter how authoritative AI outputs appear.

Guidance for practice:

• Cross-check AI-generated claims against credible scholarly and primary sources.
• Prioritize verification for data, definitions, methods, and interpretations.
• Treat fluency and confidence as insufficient indicators of accuracy.

Decision Point 5: Authority and Credibility

Guiding question: Is this a reliable and appropriate source?

 

What research suggests: Students recognize that AI can be useful for summarization, language correction, and writing support while remaining uncertain about its trustworthiness. AI outputs may also contain hallucinations or lack sufficient contextual and disciplinary understanding. Credibility must therefore be established through evidence outside the AI system.

Implication for research practice: Researchers must distinguish between information that is convincing and information that is credible within a scholarly context.

Guidance for practice:

• Validate AI outputs against peer-reviewed literature and library databases.
• Confirm that sources are appropriate for your research purpose and audience.
• Use scholarly sources as the standard for credibility, not AI-generated summaries alone.

Decision Point 6: Source Verification

Guiding question: 

Can I locate the source and confirm that it supports the information attributed to it?

 

What research suggests: AI-generated research content may include inaccurate, inappropriate, incomplete, or unverifiable references. AI can also misrepresent what a real source says or combine correct and incorrect information in a convincing response. In one study of ChatGPT-4o-generated global-health papers, 45.7% of 729 references did not exist, and only 14.4% of all generated references both existed and were relevant where cited. These risks become significant when researchers accept AI-generated citations or source descriptions without consulting the original materials.

Implication for research practice: Researchers must treat AI-generated references and claims as leads for further investigation, not as verified scholarly evidence.

Guidance for practice:

• Independently locate and verify every reference before using it.

• Confirm that the source exists and that its bibliographic information is accurate.

• Read the original source to determine whether it supports the claim attributed to it.

• Use AI to assist discovery when appropriate, but rely on library databases, scholarly records, and primary sources for validation.

Purpose: Use AI to support organization, revision, and reflection while keeping your own argument, interpretation, and voice at the center.

Decision Point 7: Integration

Guiding question: How do I use AI without replacing my thinking?

 

What research suggests: Effective AI-supported writing involves purposeful and iterative engagement. Researchers may obtain stronger results when they actively question, evaluate, revise, and reintegrate AI suggestions. Treating AI as a passive information source or accepting its first response without evaluation can limit the quality and intellectual depth of the work.

Implication for research practice: The central issue is not simply how much AI is used, but whether the researcher remains actively involved in directing and evaluating the work.

Guidance for practice:

• Use an iterative process of questioning, evaluation, revision, and reflection.
• Compare AI suggestions with your own ideas and supporting evidence.
• Revise or reject outputs that do not serve your purpose.
• Do not treat an initial AI response as finished content or allow it to determine the argument for you.

Decision Point 8: Authorship and Voice

Guiding question: What is my intellectual contribution?

 

What research suggests: Frequent or passive reliance on AI can diminish authorial control and make writing more formulaic. Nevertheless, deliberate collaboration with AI can support the writing process without eliminating the researcher’s voice or agency. The outcome depends on whether the researcher continues to control the ideas, interpretations, decisions, and revisions.

Implication for research practice: Researchers must ensure that their intellectual contribution remains visible in the selection of questions, interpretation of evidence, construction of arguments, and expression of conclusions.

Guidance for practice:

• Develop the central ideas and claims through your own reasoning.
• Use AI to test, clarify, organize, or refine those ideas, not to determine the substance of the work.
• Review AI-assisted passages for generic language or structures that obscure your perspective.
• Ensure that the final explanation, interpretation, and argument reflect your judgment and voice.