Show a desktop notification when the AI TA finishes replying
Notify me when classmates post messages in the forum
Play an alert sound whenever there is a new notification
Uedu gives every course its own AI teaching assistant — one the instructor designs, with Socratic dialogue that guides students toward thinking rather than handing them answers. Every learner–AI exchange is analyzed along Bloom's taxonomy, so instructors can see not just how much students use AI, but what kinds of thinking they engage in.
Uedu (uedu.tw) is an AI-assisted learning platform developed and operated by Chia-Kai Chang, Assistant Professor at National Central University, Taiwan, and adopted by instructors across Taiwanese universities. It runs entirely in the browser — nothing to install, nothing to deploy on your campus — and it is free for educators, with no commercial model behind it.
The platform is also a research instrument. Each learner–AI exchange is stored as a time-stamped, course-situated event carrying the dialogue, its cognitive-level classification, and its linguistic features — the data resolution behind our peer-reviewed findings on how interaction design shapes learning.
Designed for instructors in any discipline — programming, economics, statistics, languages, and more. No coding required.
Write the persona and behavior guidelines for your course's AI teaching assistant in plain language, and share it with your class through a course code.
Set discussion topics for your own material. The AI guides students with questions instead of answers, steering conversations toward higher-order thinking.
Dialogue is classified along Bloom's taxonomy, giving you a semester-long view of the cognitive levels your students engage in — not just usage counts.
AI-generated quizzes, worksheets, discussion forums, and classroom recording tools — a formative-assessment toolkit around the AI assistant.
Every major feature is backed by published research conducted on the platform itself. Three recent results:
Uedu operationalizes multimodal learning analytics in everyday courses through the Educational Omics framework: each course generates parallel, time-aligned layers of learning data. Three layers come from platform interaction alone — no sensors, no extra instruments, no added burden on students.
Cognitive processes, operationalized from dialogue: every conversational turn is classified along Bloom's taxonomy, from Remember to Create.
The language of learning: prompt specificity, semantic complexity, and how students' question-asking matures over a semester.
Social interaction: forum discussions, peer replies, and how ideas and solutions propagate through a class.
Research data collection is opt-in under an IRB-approved protocol, with consent separate from platform use.
Research participation never affects course grades. Students who decline or withdraw keep full access to the platform and their courses.
De-identification follows a documented SOP; cross-border sharing is limited to de-identified derived data under data-use agreements.
For international cohorts, student data can be stored on in-region infrastructure to meet partner-institution requirements.
Adopt the platform in your own courses and studies, independently — we ask only that publications cite the platform. See independent research built on Uedu.
Because the platform is self-built, interventions can be implemented to your design — dialogue strategies, feedback mechanisms, condition assignment. Roles and authorship are agreed before work begins.
Whether you want to try Uedu in a course, explore the data for a research question, or co-design a study, the first step is the same — write to us. We usually reply within a few days.
Chia-Kai Chang · Assistant Professor, Center for General Education,
National Central University, Taiwan
[email protected] ·
chia-kai-chang.github.io ·
eduomics.org