AI in Higher Education: Ethics, Pedagogy, and Policy

Elicit

• What it does: Searches scholarly databases to find relevant research papers, summarizes key findings, and can extract structured data directly from studies.
• Best use cases: Quickly scanning the literature for new topics, identifying gaps in research, compiling datasets for meta-analyses.
• Why it matters: Saves hours of manual reading during the early stages of a literature review.

ResearchRabbit

• What it does: Creates visual “maps” of research networks starting from a few key papers. You can explore related authors, themes, and citations in an interactive graph.
• Best use cases: Discovering hidden connections in a research area, identifying potential collaborators, and exploring interdisciplinary angles.
• Why it matters: Encourages serendipitous discovery by revealing influential works outside your usual search paths.

Scite

• What it does: Goes beyond citation counts by categorizing citations as supporting, contrasting, or mentioning the original study.
• Best use cases: Assessing the scholarly impact and reception of a paper, identifying debates or controversies in the literature.
• Why it matters: Helps you critically evaluate the strength of evidence in your field.

Consensus

• What it does: Answers natural-language questions by pulling direct findings from peer-reviewed papers.
• Best use cases: Getting quick, evidence-based overviews of a question before diving deeper into the literature.
• Why it matters: Reduces “search fatigue” and focuses your reading on the most relevant sources.

Notebook LM

• What it does: A Google AI-powered research notebook that lets you upload your own PDFs, Google Docs, and text files. It creates a personalized AI “research assistant” that can answer questions, summarize content, and generate outlines based only on the materials you’ve provided.
• Best use cases: Deep engagement with a specific set of readings, drafting literature review sections from curated sources, or preparing class materials based on selected research.
• Why it matters: Keeps your workflow private and focused on your chosen content, reducing irrelevant results and ensuring alignment with your research scope.

Scenario:
Dr. Hatch, a public health faculty member, is preparing a grant proposal on the impact of urban green spaces on adolescent mental health. She wants to quickly understand the existing research, identify gaps, and position her proposal to address an unexplored niche.

Step 1 – Deep Dive into Curated Sources with Notebook LM

• Dr. Hatch uploads 20 carefully chosen journal articles, reports, and her preliminary notes into Notebook LM.
• The tool summarizes each article, extracts common variables studied (e.g., park size, frequency of use), and helps her generate a comparative chart of methodologies.
• Because Notebook LM is working only with her curated set of documents, the summaries stay tightly focused on her scope.

Step 2 – Mapping the Research Landscape with ResearchRabbit

• Using the most influential papers identified in Notebook LM, Dr. Hatch seeds ResearchRabbit to visualize related literature.
• She spots several emerging authors and a cluster of studies in Europe that have not been cited in North American research—a potential gap for her proposal.
• ResearchRabbit’s visual network helps her identify key journals and interdisciplinary connections with environmental psychology.

Step 3 – Evaluating Evidence Strength with Scite

• Dr. Hatch inputs her top five foundational papers into Scite to see how each has been cited.
• She learns that two frequently cited studies are heavily contrasted by later research—an important nuance for her literature review.
• This insight allows her to acknowledge scholarly debates in her field, strengthening the credibility of her grant proposal.

Outcome:
By integrating Notebook LM, ResearchRabbit, and Scite, Dr. Hatch:

• Reduced her initial reading load by 40%.
• Identified an underexplored research gap.
• Strengthened her proposal’s literature review with a nuanced understanding of the evidence landscape.

Why this works:
Each tool serves a distinct role:

• Notebook LM → Deep, private engagement with curated materials.
• ResearchRabbit → Big-picture discovery of connections and gaps.
• Scite → Evidence-based citation context for critical evaluation.