PHASE ONE
Discovery
Explore a topic, develop questions, identify concepts, locate possible sources, and map the intellectual landscape.
Research activities: surveying, excavating, sorting, and mapping.
AI in Scholarly Research
AI can accelerate discovery, synthesis, and drafting—but it does not replace scholarly judgment. The Archaeological Dig Model illustrates how AI supports research while preserving human interpretation and authorship.
AI-INTEGRATED SCHOLARSHIP
The Archaeological Dig Model of AI-Integrated Research
A process model for discovery, verification, and synthesis
The Archaeological Dig Model describes research in an AI-mediated environment through three broad phases: discovery, verification, and synthesis.
AI may support work throughout the process, but researchers remain responsible for purpose, evidence selection, verification, interpretation, transparency, and scholarly authorship.
The process remains. The tools are becoming more fluid.
The model describes enduring forms of scholarly activity, not fixed categories for AI products. Many contemporary tools can support several parts of the research process.
Tools should be evaluated by their capabilities, sources, limitations, privacy practices, and influence on scholarly work, rather than assigned permanently to a single stage.
The Research Process
DISCOVERY
VERIFICATION
SYNTHESIS
Surveying
the Site
the Site
Excavating
& Sorting
& Sorting
Mapping
the Landscape
the Landscape
Authentication
& Context
& Context
Library
Authority
Authority
Synthesis &
Interpretation
Interpretation
Scholar as
Final Authority
Final Authority
The Archaeological Dig Model of AI-Integrated Research
A process model of scholarly research in an AI-mediated environment
Three Phases of Research
PHASE TWO
Verification
Confirm that sources exist, evaluate authority and context, retrieve full text, and establish a credible evidence base.
Research activities: authentication, contextual evaluation, and library-supported access.
PHASE THREE
Synthesis
Compare evidence, interpret patterns, construct arguments, communicate findings, and take responsibility for final claims.
Research activities: synthesis, interpretation, writing, and authorship.
COMPANION FRAMEWORK
Scholarly Decision Points
The Archaeological Dig Model describes the stages of research. The companion Scholarly Decision Points framework examines the recurring judgments researchers make while moving through discovery, verification, and synthesis.
Explore the Scholarly Decision PointsTOOLS WITHIN THE PROCESS
Representative Tools Across the Research Process
Many tools support more than one phase. The examples below illustrate common uses rather than fixed categories, institutional approval, or endorsement.
DISCOVERY
Explore and Map
Explore topics, identify terminology, locate scholarly literature, and trace relationships among publications.
Representative tools:
Consensus, Semantic Scholar, ResearchRabbit, Connected Papers, Litmaps, and OpenAlex
VERIFICATION
Confirm and Evaluate
Confirm sources, inspect publication information, retrieve full text, trace citations, and evaluate authority and context.
Representative tools:
Library databases, OpenAlex, Semantic Scholar, Consensus, Elicit, and Zotero
SYNTHESIS
Compare and Construct
Compare evidence, organize themes, develop interpretations, construct arguments, and communicate findings.
Representative tools:
NotebookLM, Elicit, ChatGPT, Microsoft Copilot, Consensus, and Zotero
Need task-specific tool guidance?
The AI Research Hub organizes tools by scholarly activity and provides additional guidance for choosing and using them.
Ethical, Transparent, and Reproducible AI Use
Use AI in ways that align with scholarly, disciplinary, publisher, funder, instructor, and institutional expectations.
Responsible AI Tool Use at UW Oshkosh
This page includes both campus-vetted tools and public tools shared for awareness and optional exploration.
Do not enter confidential, protected, personally identifiable, unpublished, proprietary, student, grant-review, peer-review, or restricted research information into public AI systems.
Humans remain responsible for authorship
Researchers retain responsibility for accuracy, originality, interpretation, evidence selection, claims, conclusions, and compliance with relevant policies.
Protect unpublished and confidential work
Do not upload confidential manuscripts, proposals, student work, peer reviews, unpublished data, proprietary material, or IRB-sensitive content into public systems.
Disclose meaningful AI assistance
Disclosure may be appropriate when AI contributes materially to research design, evidence processing, coding, analysis, image production, drafting, rewriting, or synthesis.
Consult the current policy of the relevant journal, publisher, professional association, funder, department, or instructor.
Example AI use statement
“NotebookLM was used to compare themes within a researcher-selected collection of published articles. All source selection, interpretation, verification, and final writing were completed by the author.”
Library Databases, Zotero, and Librarian Support
AI tools are most useful when integrated with the scholarly infrastructure that supports verified evidence, documented methods, and responsible interpretation.
Library Databases
Provide disciplinary indexing, controlled vocabulary, precise filters, licensed full text, and reproducible search environments.
Zotero
Helps transform candidate sources into an organized, annotated, verified, and citable research collection.
Librarian Expertise
Supports search strategy, evidence evaluation, citation verification, tool selection, documentation, and responsible research workflows.
A complete scholarly research ecosystem
AI supports exploration and processing → library systems establish access and authority → Zotero organizes and documents evidence → librarians strengthen strategy and evaluation → researchers provide interpretation, judgment, and authorship.
Access, Privacy, and Changing Tool Features
Some tools listed on this page are not institutionally provided and may operate through free, freemium, trial, or paid access models. Use of public tools is optional.
- Review current access terms before beginning a substantial project.
- Do not assume that a free feature will remain available.
- Review privacy, retention, training-data, and account settings.
- Do not upload protected or confidential information into public tools.
- Record the tool, version, date, purpose, and scope of meaningful use.
Maintenance note: Tool examples are representative rather than comprehensive. The research phases provide the stable structure; tool descriptions should be reviewed periodically as capabilities and access models change.
AI can support every phase. It cannot assume scholarly responsibility.
AI may contribute speed, scale, pattern recognition, and exploratory support. Researchers provide purpose, disciplinary knowledge, source evaluation, verification, interpretation, transparency, judgment, and authorship.
Researching with AI?
Polk Library can help with AI-informed search strategy, literature review planning, tool evaluation, database searching, Zotero, source verification, and transparent research workflows. Request a research consultation →
