AI at SITE 2026:
Mapping the Research Landscape

What is the field talking about when it talks about AI in education? We reviewed 102 proposal abstracts to find out.

The Big Picture
102 Papers, Four Dimensions
We reviewed the abstract of every proposal accepted by the AI SIG in the SITE 2026 conference, classifying each along four dimensions: what topic it addresses, what context it studies, what role AI plays in the research, and whether specific subject-matter content is involved. Here's what the landscape looks like.
102
Papers
7
Topic Areas
5
Roles of AI
5
Contexts
27
Name a
Content Area
The Topics
What Are We Talking About?
Seven topic areas emerged from the analysis. The conversation is dominated by two concerns: how ready educators are for AI (Competencies, Perceptions, and Adoption) and what the ethical guardrails should be (AI Ethics and Policy). Together these account for more than half of all papers. Hover over any card to see its papers.

Topic Distribution

Number of papers per topic area

The largest cluster — Competencies, Perceptions, and Adoption (34 papers) — is asking a fundamentally pre-adoption question: are we ready? The second-largest — AI Ethics and Policy (22 papers) — is asking a governance question: how do we keep this safe? Meanwhile, only 7 papers focus primarily on how AI shapes student learning itself.

The Role of AI
How Is AI Positioned?
Beyond topic, each paper positions AI differently. Is AI the tool being used? The subject being studied? A collaborator in the research process? The answer reveals how the field conceptualizes the relationship between educators and AI.

Roles at a Glance

Each square is one paper, colored by the role AI plays

Half of all papers (51) position AI as a Tool or Assistant — something to be used. A third (33) treat AI as a Subject of Inquiry — something to be studied and understood. Only 10 position AI as a Partner or Collaborator, and just 4 as a Tutor, Coach, or Generator. The field is still in an early phase: figuring out what AI is before figuring out what to do with it.

Role of AI by Topic Area

How each topic positions AI
The Context
Where Is This Research Happening?
As a teacher education conference, it's no surprise that Teacher Education is the most common research context. But note that nearly a quarter of papers don't specify a context at all — they discuss AI in education generically.

Research Context

Where the research takes place
Content Areas
Does Subject Matter Matter?
We also asked: does each paper engage with a specific subject-matter content area — math, science, literacy, the arts — or does it treat AI as content-agnostic? The vast majority discuss AI without reference to any specific discipline.

Content Area Breakdown

Each square is one paper

Only 27 of 102 papers (26%) name a specific content area. Of those, Literacy & Languages leads with 13, followed by STEM with 9. The remaining 75% discuss AI as if the subject being taught doesn't matter — a notable gap for a field that knows pedagogy is always shaped by content.

Explore
The Research Landscape
Each dot below is a paper, positioned by thematic similarity using computational text analysis. Hover for details and switch between views.
Methods
How We Did This

Data Collection & Classification

We collected all 102 unique propsals accepted to by the AI SIG for SITE 2026, extracting titles, abstracts, and keywords. Each paper was classified along four dimensions: Topic Area (the primary research focus), Context (the educational setting studied), Role of AI (how AI is positioned in the paper), and Content Area (whether specific subject matter is involved).

Visualization

The topic landscape uses TF-IDF vectorization of paper text, reduced to two dimensions via t-SNE. The keyword network shows co-occurrence relationships among the most frequent terms. All interactive visualizations are built with D3.js.

A Note on AI-Assisted Analysis

This analysis was conducted with the assistance of Claude (Anthropic). Metadata extraction, thematic clustering, and visualization generation were AI-assisted. Classifications were reviewed and corrected by the researcher, but some inaccuracies may still exist.