Conversation Tasks

💬 What is a Conversation Task in Classlet?

Conversation tasks in Classlet are dynamic, agent-based interactions where learners engage in spoken or typed exchanges with GPT-powered avatars or scripted characters. These tasks are implemented across both desktop/mobile and VR modes — in either 2D chat or 3D avatar formats.

These conversations are not just dialogue boxes — they are pedagogical encounters that allow learners to receive instruction, simulate roles, and practice reasoning or language in context

This dialogue in a scene for Lingnan University, introduces learners to 1930s New York during the Great Depression, using a narrative-driven NPC to set historical context. Students are immersed in a socio-economic environment marked by poverty and unemployment, priming them for critical thinking and decision-making in a historically grounded scenario.

🎯 Key Pedagogical Purposes

Function	Description
Knowledge Delivery	Agent provides structured, chunked explanations during conversation
Feedback & Remediation	Agent offers correct/incorrect response branches with built-in guidance
Scenario Simulation	Learner engages in social or procedural talk (e.g., dentist/patient roles)
Practice & Repetition	Learner repeats key vocabulary or logic steps across conversation paths

🎭 Role-Based Interaction

In VR or mobile/desktop formats, agents are assigned roles such as:

• Instructor – delivers new knowledge, theories, or procedures.

• Mentor/Coach – supports reflection and encourages metacognition.

• Challenger – poses dilemmas or provocative follow-ups.

• Peer/Patient/Client – engages learner in simulated real-world roles (e.g., ethics, nursing).

This structure aligns with situated learning theory, where knowledge is embedded in social and task-specific contexts (Lave & Wenger, 1991).

📚 Grounding Theories

• Cognitive Apprenticeship: Learners observe modeled language use and then participate within safe, guided scaffolding (Collins et al., 1989).

• Situated Learning: Knowledge is embedded in socially and contextually meaningful interactions (Lave & Wenger, 1991).

• Vygotskian Dialogue: Dialogue fosters higher-order reasoning when scaffolded within a learner’s zone of proximal development (Vygotsky, 1978).

• Instructional Alignment: Branches ensure responses reinforce target content, avoiding off-topic drift (Biggs & Tang, 2011).

📐 Design Considerations

• Use simple response trees for fast tasks; deeper trees for storytelling or dilemmas.

• Include positive, neutral, and incorrect reply paths.

• Tie response branches directly to the target learning objective (cognitive, procedural, affective).

• Reuse key phrases for reinforcement and vocabulary exposure.