What work did AI perform?
Identify whether AI supplied answers, produced material, explained concepts, or helped solve a problem alongside you.
AI engagement, made visible
See how you are thinking with AI.
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
Conversation patternIllustrative sample, not a score
“Summarize these three studies into the table format I specified, and flag any claim that needs a source.”
What the analysis reveals
Not whether you used AI. How.
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.
What work remained yours?
See where you set the problem, challenged assumptions, verified claims, expanded possibilities, and made consequential choices.
The eight-mode framework
A vocabulary for seeing the difference.
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.
Agency
The person directs, evaluates, reframes, or tests the work.
Partnership
The person and AI exchange explanations, questions, and partial solutions.
Passivity
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
Build range, not dependence.
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
Make engagement assessable.
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
Take the model with you.
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
Student Guide to AI-Assisted Learning
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
Faculty Guide to AI-Assisted Research
A research-workflow companion for literature, methods, analysis, drafting, and revision, with retained scholarly judgment and a research-specific brief.
For Partners and researchers
AI Engagement Model — Overview Deck
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
A working instrument, grounded in an evolving model.
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.
One model, several ways in
Use the site by the job you need to do.
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.