AI @ UA: For Researchers

Image description
Whether you're just getting started or already building AI into your workflow, these resources are designed to help you use these tools effectively, ethically, and in ways that strengthen rather than shortcut your scholarship.

 

Where Are You Starting From?

Start with Know Your Tools and NotebookLM
Jump to Literature Reviews, Grant Writing, or Citation & Disclosure
See AI for Coding and Data Science

 

Getting Started: AI Basics for Researchers

  • What these tools are and aren't
  • The essential habit: verify everything, especially citations and factual claims
  • UA-approved tools (Gemini, NotebookLM, Copilot) and why that matters for data security
  • Discipline-specific norms vary—your professional association and journal guidelines take precedence

 


Spotlight: Getting Started with NotebookLM

  • What makes NotebookLM different: it works only from documents you upload, dramatically reducing hallucination risk
  • Use cases: querying across multiple PDFs, generating summaries from field notes or transcripts, creating briefing documents, audio overviews
  • Key caveat: a tool for working with texts you've gathered, not a substitute for reading them
  • Links to: Google's official NotebookLM guide; academic reviews of NotebookLM in research workflows

AI for Grant Seeking and Grant Writing

Finding Funding

  • AI can match your research profile to open solicitations and analyze RFPs for alignment with your work
  • Recommended platforms: Pivot-RP, Grants.gov, NIH Reporter, NSF Award Search

 

Drafting and Refining Proposals

  • Best uses: early-stage drafting of specific sections (broader impacts, project summaries, literature context), checking that all RFP requirements are addressed, adjusting reading level for non-specialist reviewers
  • Important caveat: requires heavy human revision to reflect your actual research vision

 

Budgets, Logic Models, and Gantt Charts

  • AI can generate first-draft budget narratives and flag common line-item categories for your project type—but your sponsored programs office must review all figures
  • Logic models: AI is useful for drafting the input/output/outcome structure and stress-testing the logical connections between program activities and goals
  • Gantt charts: AI can translate a narrative project description into a draft timeline; tools like Copilot integrate directly with Excel for formatted output
  • In all three cases: AI produces a useful scaffold, not a finished product

 

Disclosure and Compliance

  • NIH, NSF, and other major funders are developing AI disclosure requirements—check current guidance before submitting
  • Links to: NIH grants guidance; NSF AI policy; NCURA resources; your UA Office of Research and Sponsored Programs

 


AI and the Literature Review: A Genuine Debate


The case for AI-assisted lit reviews

  • Surfaces relevant papers from large corpora, identifies patterns, flags contradictions across studies.
  • Surfaces strong work buried in poor writing; exposes weak work propped up by confident prose.

 

The case for caution

  • You cannot fully absorb a text or engage with its argument unless you sit with it slowly; speed-reading a summary is a different cognitive experience than wrestling with a difficult paper. 
  • Some ideas only crystallize through prolonged immersion. The researcher who has genuinely read 200 papers thinks differently than one who has processed 2,000 summaries. 

 

Concrete limitations

  • Frequent hallucination of citations: plausible-sounding but nonexistent papers, or misattributed findings.
  • Limited access to recent scholarship, proprietary databases, or subfield-specific context.

 

A practical tiered approach

  • Use AI freely: identifying search terms, scanning abstracts, formatting citations, checking argument logic.
  • Use AI carefully and verify: synthesizing findings, drafting background sections
  • Don't outsource to AI : close reading of foundational texts, evaluating individual study quality, forming your scholarly argument.
  • Look to disciplinary associations and target journals for field-specific norms.

 


AI for Academic Writing: What the Research Community Is Actually Saying

  • Widespread private use outpaces public acknowledgment, a gap corrosive to norms and trust.
  • Strongest legitimate uses: editing and proofreading, restructuring arguments, reformatting for new audiences, boilerplate sections (data availability statements, IRB summaries).
  • Core concern: AI writing tends toward confident vagueness. It may be authoritative sounding but lacking the precise intellectual commitments that define good scholarship.
  • For new researchers especially: use AI to help you write better, not to write for you. Your scholarly voice is an asset you're building over a career

 


AI for Coding and Data Science

Tools worth knowing

  • GitHub Copilot: real-time code suggestions in VS Code and other IDEs; check with UA IT for university licensing.
  • Claude Code: command-line coding agent for writing, editing, debugging, and documenting across full projects.
  • OpenAI Codex: underlying model for many coding assistants; accessible via API for custom integrations.

 

Highest-value uses for researchers

  • Boilerplate for data cleaning, reformatting, and pipeline setup.
  • Debugging and explaining unfamiliar code.
  • Translating between languages (R to Python, etc.).
  • Documenting existing code for reproducibility.

 

A note on agentic tools

  • Agents can now execute multi-step data workflows autonomously; a rapidly developing area.
  • Key caution: a plausible-looking result is not the same as a correct one. Always validate outputs against your own understanding of the data
  • Links to: GitHub Copilot for Education; fast.ai; relevant professional associations (ASA, ACM) for field-specific guidance

 


Citing and Disclosing AI Use

  • Why it matters: readers and reviewers need to know how work was produced to evaluate it fairly.
  • APA 7th: acknowledge in text, cite tool as non-retrievable sources: apastyle.apa.org
  • MLA: Works Cited entry plus in-text note explaining use: style.mla.org
  • Chicago: acknowledgment in footnote or author's note:  chicagomanualofstyle.org
  • Journal policies vary widely: full prohibition / editing only with disclosure / required disclosure / no policy yet—always check author guidelines; when ambiguous, disclose anyway
  • Links to: COPE guidelines on AI authorship; Nature's AI disclosure policy as a model

 


How AI Is Being Used in UA Research

  • Spotlight: UA Researchers using AI across disciplines
    • Dr. Anup Kumar Das (UAF Data/AI Lab)

      At the University of Alaska Fairbanks (UAF), Dr. Anup Kumar Das leads the Data/AI Lab within the Computer Science Department. He recently spearheaded a major high-performance computing upgrade to UAF’s Chinook supercomputer cluster, introducing advanced NVIDIA H100 GPUs specifically to handle complex generative AI, deep learning, and big data problems. 

      Published Focus: His lab works across multiple science and engineering disciplines, training models on immense datasets related to Alaska’s distinctive climate, flora, and fauna. 
      For more information on his current projects, you can look into the UAF Data/AI Lab.
    • Dr. Kenrick Mock & ADSAIL (UAA)

      At the University of Alaska Anchorage (UAA), Dr. Kenrick Mock (Dean of the College of Engineering) helped establish ADSAIL (Alaska's Data Science & Artificial Intelligence Lab). This interdisciplinary lab acts as a collaborative hub for all UA campuses.

      Published Focus: Dr. Mock’s published research interests cover machine learning, computer vision, and complex systems. He works alongside other UAA faculty like Dr. Mohammad Heidari Kapourchali, who publishes on utilizing machine learning and signal processing to optimize cyber-enabled systems and power grids in harsh environments.

      Learn more via the UAA Artificial Intelligence and Robotics (AIR) Lab

  • Emphasis on Arctic research, Indigenous data sovereignty, and Alaska-specific applications
  • Invite submissions from faculty and graduate researchers

 


Guidelines, Policies, and Where to Get Help

  • UA sensitive data guidance: FERPA, HIPAA, IRB considerations when using AI with research data
  • Links to UA policy and federal agency guidance (NIH, NSF, NEH)
  • Office of Research and Sponsored Programs contact
  • UA Library resources for AI-assisted research

 

  1. UAS Egan Library: Artificial Intelligence Guide

The University of Alaska Southeast (UAS) Egan Library hosts a dedicated, centralized hub for researchers navigating the world of machine learning and large language models.

  • What it offers: A curated collection of resources designed to increase knowledge of artificial intelligence. It features dedicated guidance on AI prompting strategies specifically optimized for academic exploration, a breakdown of UA-sanctioned AI tools, and instruction on navigating built-in AI features hidden inside standard library databases.
  • Access the portal directly via the University of Alaska Southeast Faculty Services (listed under the Artificial Intelligence Guide resource).

 

  1. UAF Elmer E. Rasmuson Library & Center for Teaching and Learning (CTL)

At the University of Alaska Fairbanks (UAF), the library and CTL collaborate closely to guide researchers on utilizing AI as an analytical writing and data-sorting assistant.

  • What it offers: Resources that instruct researchers on how to use generative AI for higher-order academic tasks. This includes polishing language, manipulating massive datasets/spreadsheets, solving complex research equations, and identifying blind spots or opposing viewpoints in literature reviews. The library provides clear orientation on the thin line between using AI as a legitimate research mechanism versus accidental plagiarism.
  • Read the guidance frameworks at the UAF Center for Teaching and Learning AI in Education Guide.

 

  1. UAA/APU Consortium Library: Academic Integrity & AI

The University of Alaska Anchorage (UAA) and Alaska Pacific University (APU) Consortium Library integrates AI into its fundamental research assistance and self-guided help sections.

  • What it offers: Through their Academic Integrity & AI initiative, the library helps researchers understand how to properly document, credit, and cite AI-assisted discovery methods (such as citing code written by AI or text parsed by LLMs). It maps out how to responsibly use AI tools alongside traditional citation software.
  • Explore their research workflows on the UAA/APU Consortium Library Research Help Portal.
  • When in doubt, ask a librarian—research librarians are among the most knowledgeable people on campus about AI tools for scholarship