AI/ML research
Student Workflow Interviews
Berkeley, CA
Apr 14, 2026
A focused research pass on how students bring ChatGPT into studying, assignments, planning, and feedback loops.
Students do not use AI in one clean flow. They jump between asking, checking, rewriting, dodging friction, and trying not to lose ownership of the work.
I interviewed students across K-12 and college, looked for repeated friction points, and grouped the findings around moments where interface design could help.
It gave the ChatGPT Lab project a grounded view of student workflow needs instead of generic AI-education assumptions.
It supports sharper product questions: when should AI ask, explain, slow down, summarize, or hand control back to the student?
Highlights
- Used interviews to map student AI workflow friction.
- Separated convenience use cases from deeper learning moments.
- Converted qualitative patterns into concrete product design questions.