Artefacts
Research software I build to study the work, rather than ask about it afterwards.
A trace-capturing co-creation environmentIn development
The research instrument for my doctorate: a web environment where a teacher educator and a generative AI model co-create teaching material while every move either party makes is recorded. Prompts, suggestions, accepts, rejects, edits, reversions, pauses.
Most studies of AI in professional work ask people afterwards what happened. Self-report is exactly where judgment goes invisible; you do not remember the suggestion you waved through. Capturing the interaction itself lets me see where a professional overrode the model, where they went along with it, and what they made of what came back, rather than what they later recall doing.
TypeScript · Vite · trace logging · TPACK-informed task design
Built under a research-software discipline: reproducible, accessible, and with the empirical data boundary enforced in code rather than in good intentions.
Code: graite-instrument on GitHub.