ClassAid: A Real-time Instructor-AI-Student Orchestration System for Classroom Programming Activities

TL;DR

ClassAid presents a real-time instructor-AI-student orchestration system that integrates TA Agents with an instructor dashboard to provide personalized feedback in classroom programming activities. Rooted in formative and dynamic assessment theories, the authors implement a six-stage TA Agent framework that observes, diagnoses, and intervenes to support metacognitive development while preserving teacher authority. A classroom deployment with 54 students and follow-up educator interviews demonstrate high agent accuracy, useful real-time insights, and positive reception, though they also reveal challenges related to latency, scalability, and Auto-mode decisions. The work offers design implications for interactive AI in authentic classrooms and outlines future directions toward finer-grained feedback, shared control, and broader generalizability across contexts.

Abstract

Generative AI is reshaping education, but it also raises concerns about instability and overreliance. In programming classrooms, we aim to leverage its feedback capabilities while reinforcing the educator's role in guiding student-AI interactions. We developed ClassAid, a real-time orchestration system that integrates TA Agents to provide personalized support and an AI-driven dashboard that visualizes student-AI interactions, enabling instructors to dynamically adjust TA Agent modes. Instructors can configure the Agent to provide technical feedback (direct coding solutions), heuristic feedback (hint-based guidance), automatic feedback (autonomously selecting technical or heuristic support), or silent operation (no AI support). We evaluated ClassAid through three aspects: (1) the TA Agents' performance, (2) feedback from 54 students and one instructor during a classroom deployment, and (3) interviews with eight educators. Results demonstrate that dynamic instructor control over AI supports effective real-time personalized feedback and provides design implications for integrating AI into authentic educational settings.

Figures (5)

• Figure 1: Student interface during in-class programming activities. (A) The task panel shows the current feedback mode (a1), task description (a2), data (a3), and expected output (a4). (B) The chat panel supports student questions and TA Agent responses, with options to rate messages and receive proactive feedback (b1). (C) The code panel allows students to write and run code. (D) The output panel displays execution results and error messages.
• Figure 2: Overview of the TA Agent's six-stage orchestration pipeline for student learning support within the ClassAid system. Student activity data, such as question submissions, code execution, and interaction traces, are first collected through the ClassAid Student Interface and passed to the TA Agent backend. The agent then enters a six-stage pipeline that observes, analyzes, and responds to student learning behaviors. Meanwhile, instructors monitor student progress and the TA Agent's performance through the ClassAid Instructor Dashboard and can adjust feedback modes in real time to align with pedagogical goals.
• Figure 3: Instructor dashboard for real-time classroom orchestration. (A) Class-Level Alerts highlight potential learning risks through Agent, Process, and Outcome alerts. (B) Class-Level Analysis aggregates question (b1) and code (b2) issues to reveal class-wide bottlenecks. (C) Student Performance Cards display each student's task score and current feedback mode, with global controls for mode switching (c3). (D) More Details Panel provides drill-down views of individual students, including agent interactions (d1), mode control (d2), and task-level analysis (d3). Together, these components enable timely intervention and data-informed teaching decisions.
• Figure 4: The Code Analysis view shows how often different code problems appear in student submissions. Instructors can click on each bar to see which students had that issue and and how many times it occurred.
• Figure 5: Student ratings on the ClassAid.