AI engagement, made visible
Analyze a real conversation to see when you delegated judgment, worked in partnership, or exercised intellectual agency. Then use the result to choose your next move.
Eight observable engagement modes · Three flexible groupings · Research-backed feedback
“Summarize these three studies into the table format I specified, and flag any claim that needs a source.”
45–60 minute certificate activity
Complete an interactive lesson, practice in two AI conversations, and demonstrate your understanding. Earn a certificate when you meet the requirements.
Start the certificate activity Preview the theory lessonFor instructors
Use the introductory certificate activity, create assignments, and follow your students’ engagement results. Students join with a course code. No MyEducator book required.
Explore instructor tools Create accountWhat the analysis reveals
The useful question is not simply whether AI was present. It is how responsibility moved between you and the system, and whether your own reasoning remained visible.
Identify whether AI supplied answers, produced material, explained concepts, or helped solve a problem alongside you.
See where you set the problem, challenged assumptions, verified claims, expanded possibilities, and made consequential choices.
The eight-mode framework
The modes describe observable functions in a conversation. They organize the many ways people work with AI without reducing every interaction to “good” or “bad.”
A mode is an observable function, not an identity or a moral label. Its value depends on the task, the timing, the stakes, and the control you keep. The colours distinguish the three groupings; they do not rank them.
The person directs, evaluates, reframes, or tests the work.
The person and AI exchange explanations, questions, and partial solutions.
AI supplies or produces; the person's reasoning may remain hidden.
Display convention: Agency appears above Partnership and Passivity. These are flexible patterns within tasks and moments, not identities, developmental levels, or a hierarchy of people. A person can move among them, and any mode can be wise or unwise depending on context.
For learners
Practice moving deliberately among the modes. Use AI as a tutor when you need explanation, a collaborator when you need exchange, and a target for verification when accuracy matters.
Start guided practiceFor instructors
Give students a shared language for documenting how they used AI, then inspect conversation evidence instead of relying only on declarations or final outputs.
Explore instructor toolsGuides and toolkits
Three downloadable companions carry the full framework into study, research, and briefing work. The editable versions are there to be adapted for a course.
For Students
A practical handbook for learning with the complete model: the formation tree, eight modes, a five-part prompting brief, and copy-ready study workflows.
For Faculty
A research-workflow companion for literature, methods, analysis, drafting, and revision, with retained scholarly judgment and a research-specific brief.
For Partners and researchers
A presentation companion for partner briefings and faculty workshops. Pair it with the formation tree and the interactive framework page.
Plugin files are Claude Code toolkit bundles. Drop one into your plugins directory to install its skills; each archive carries its own install notes.
Research in progress
AI Modes connects an eight-mode typology, research methods, teaching practices, and ongoing studies. Methodology, evidence, limitations, and publications belong together, not behind a marketing claim.
Built by Mark Keith, PhD, at Brigham Young University.
One model, several ways in
The structure stays compact: analyze a conversation, learn the practice yourself, teach it to a class, or inspect the evidence underneath. Everything deeper sits one level in, without crowding the first decision.