Polk Library

AI-Supported Science Research Workflow (Biology Example)

AI-Supported Science Research Workflow: Biology Example

Goal: Help students use AI productively when developing research papers or other science-based assignments.

This workflow is designed to support student use of AI in science-based research assignments and can be adapted across biology topics and course levels. This workflow is adapted from the AI-Integrated Research Archaeological Dig Model and demonstrates how AI tools can support scientific research processes. Additional resources are also located on our AI Research Hub.

 

✅ Using AI in Research (Core Guidance)

This guide may include both campus-supported tools and public AI tools to support research, topic exploration, and learning.

According to Universities of Wisconsin guidance on AI tools (PDF) , students should use AI thoughtfully, verify information

carefully, and follow course expectations.

 

Privacy reminder: Do not enter confidential, sensitive, or personally identifiable information (PII) into any AI tool. This includes

student records, unpublished work, or restricted materials.

 

Course reminder: Always follow your instructor’s expectations. No student should be required to create an account for a non-

university tool.

⚠️ Key Reminders for Academic Research

  • AI should be used as a starting point. It supports exploration—but does not replace reading and analyzing sources.
  • Claims must be supported by evidence. Use peer-reviewed or authoritative sources.
  • All information must be verified. AI outputs may be incomplete or inaccurate—confirm using library databases.
  • Use AI in accordance with course and university expectations.
  • You are responsible for your work. You—not the AI—are accountable for accuracy and conclusions.
When and How to Cite AI

Cite AI if:

  • You include AI-generated text (quoted or paraphrased)
  • You rely on AI-generated structure, analysis, or explanations

No citation is usually needed if:

  • You used AI only for brainstorming or idea development
  • No AI-generated content appears in your final work

Common Citation Styles

 

APA (7th ed.)

OpenAI. (2025). ChatGPT (GPT-5.3) [Large language model]. https://chat.openai.com/

Guidance: APA – Citing Generative AI

MLA (9th ed.)

OpenAI. ChatGPT, 2025, https://chat.openai.com/.

Guidance: MLA – Citing Generative AI

Chicago (17th ed.)

OpenAI. ChatGPT. 2025. https://chat.openai.com/.

Guidance: Chicago Manual of Style – AI Citation Guidance

 

Optional transparency note:

“AI tools were used for initial brainstorming and keyword development. All sources were verified and interpreted by the author.”

 

Important: Always cite the original sources that support your claims. AI is not a substitute for scholarly evidence.

Decision Points in This Workflow

As you move through the research process, pause to ask how AI is influencing your decisions.

  • Discovery: When and why should I use AI? How is AI shaping my understanding of the topic?
  • Verification: Is this information correct? Can I verify where it comes from?
  • Synthesis: How do I use AI without replacing my thinking? What is my intellectual contribution?
When and How to Cite AI

Cite AI if:

  • You include AI-generated text, whether quoted or paraphrased.
  • You rely on AI-generated structure, analysis, or explanations.

No citation is usually needed if:

  • You used AI only for brainstorming or idea development.
  • No AI-generated content appears in your final work.

Synthesis decision point:

Authorship and Voice: What is my intellectual contribution? Make sure your final work reflects your own analysis, interpretation, and responsibility.

Common Citation Styles

APA, 7th ed.

OpenAI. (2025). ChatGPT (GPT-5.3) [Large language model]. https://chat.openai.com/

Guidance: APA: Citing Generative AI

MLA, 9th ed.

OpenAI. ChatGPT, 2025, https://chat.openai.com/.

Guidance: MLA: Citing Generative AI

Chicago, 17th ed.

OpenAI. ChatGPT. 2025. https://chat.openai.com/.

Guidance: Chicago Manual of Style: AI Citation Guidance

Optional transparency note:

“AI tools were used for initial brainstorming and keyword development. All sources were verified and interpreted by the author.”

Important: Always cite the original sources that support your claims. AI is not a substitute for scholarly evidence.

Step 1: Explore a Topic and Refine a Research Direction

You can begin with a topic, question, or area of interest. There are two useful ways to begin discovery in science research:

Discovery decision points for this step:

  • AI Engagement: When and why should I use AI?
  • Problem Framing: How is AI shaping my understanding of the topic?
  • Relevance and Positioning: What information is worth keeping, and how does it fit?

Option A: Literature- and citation-based discovery

Use Semantic Scholar to begin from the scholarly literature itself.

  • Identify important papers, authors, and terminology.
  • See how a topic is described in the research literature.
  • Build vocabulary for database searching.

Option B: Research-question-based discovery

Use Consensus to begin from a topic, prompt, or rough question and refine it into a researchable direction.

  • Brainstorm focused, researchable questions.
  • Identify technical vocabulary and synonyms.
  • Generate concepts and phrases to try in databases.

Try this Consensus prompt:

I am a biology student preparing a research paper. I am interested in:

[insert topic]

Suggest several focused researchable questions that:

  • Investigate a specific biological mechanism or process.
  • Are supported by peer-reviewed research.
  • Could be explored using experimental evidence.
  • Could appeal to high school, undergraduate, or graduate students.
  • Include a contentious question for each.

For each question, include keywords and search phrases for databases like PubMed or Web of Science.

Important: Both approaches are starting points. Strong research still moves into database searching and close reading of primary research.

Step 2: Search Strategically Across Databases

Regardless of how students begin, strong research moves into structured database searching.

No single database indexes all scientific literature. Choose databases based on the research angle and compare results.

Verification decision points for this step:

  • Authority and Credibility: Is this a reliable and appropriate source?
  • Source Verification: Can I locate and verify where this information comes from?
  • Accuracy: Does the source support the claim I want to make?

Major Research Databases

Discipline-Specific Journal Platforms

Best Practice:

  • Search at least two databases.
  • Compare terminology and results across platforms.
  • Use citation chaining, including “Cited by” tools and reference lists.
  • Refine your Boolean search string as vocabulary evolves.

Your goal is to locate peer-reviewed primary research articles with experimental data, not summaries or background overviews.

Not sure which database best fits the assignment or topic? Consider the research angle below.

Advanced Strategy

If a topic spans multiple biological scales, such as gene regulation, metabolism, and ecological impact, students may need to search different databases separately.

  • Molecular mechanisms → PubMed
  • Interdisciplinary or citation mapping → Web of Science
  • Ecological systems → Wildlife & Ecology Studies Worldwide
  • Metabolites or signaling molecules → ACS Journals

Strong science research often requires strategic comparison across databases.

Step 3: Evaluate Experimental Strength

Before using an article, students should confirm that it:

  • Contains a Methods section.
  • Includes original experimental data, such as figures or tables.
  • Tests a hypothesis or research question.
  • Uses appropriate controls.

Verification decision points for this step:

  • Accuracy: Are the article’s claims supported by its data?
  • Authority and Credibility: Is this article appropriate for a science-based research assignment?
  • Source Verification: Can you trace the evidence back to the original study?

Review articles are useful for background, but they should not be the main evidence for most research assignments.

Reference Management

Zotero: Save citations, attach PDFs, organize sources, and generate bibliographies.

Research responsibility reminder:

A citation manager can help organize sources, but it does not verify whether a source is accurate, relevant, or appropriate. Always check citations, article details, and evidence yourself.