IFIP TC3 OCCE 2025

Recognising Need,
Requesting Support

GenAI–Supported Socially–Shared Regulation in Online Collaborative STEM Inquiry Learning. How students actively recruit AI as a regulatory partner using the GenAIRL tool.

Ucan, S., Webb, M., & Galamba, A. (2025). Recognising Need, Requesting Support: GenAI–Supported Socially–shared Regulation in Online Collaborative STEM Inquiry Learning. IFIP TC3 OCCE 2025.

verified This research was funded by TUBITAK 2219 - International Postdoctoral Research Fellowship Program (Project No: 1059B192400202).
school Conducted while the first author, Assoc. Prof. Dr. Serkan Ucan, was a visiting scholar at University of California, Berkeley.
AI Collaboration

What is GenAIRL?

Generative AI for Regulated Learning.

A Google Docs add-on tool designed on the CHARM Framework (Collaborative Human-GenAI Regulation Model). Unlike generic chatbots, GenAIRL is context-aware and designed specifically to scaffold regulatory skills.

psychology
Context-Aware Prompting Reads document text & instructions in real-time.
group_work
Dual-Agent Model Micro-Agent: Reactive chat partner.
Macro-Agent: Periodic proactive feedback.

science Study Context

Participants: 4 groups (N=14) of undergraduate bio-engineering students.
Task: 8-week collaborative inquiry project on marine pollution in the Marmara Sea.
Data: Mixed-methods analysis of 119 AI queries, time-stamped logs, and 16 Zoom session recordings.

flag Research Goal

GenAI often risks fostering over-reliance. This study investigates the specific "Trigger Events" that prompt students to voluntarily recruit AI as a partner, and the nature of their requests.

Analytical Workflow

Zoom Transcripts arrow_forward GenAIRL Logs arrow_forward Trigger-Request Analysis

Key Research Findings

psychology

Metacognitive Drivers

74.1% of student triggers were metacognitive (planning, monitoring, evaluating) rather than simple knowledge gaps. Students use GenAIRL to regulate their work, not just do it.

handshake

The "Critical Friend"

The most frequent request (43.3%) was for Regulatory Support, specifically seeking evaluation of work and ideas. GenAIRL acted as an auditor for group progress.

radio_button_unchecked

Regulatory Loops

Highly predictable Trigger-Request patterns emerged. E.g., Internal monitoring ("Are we on track?") directly triggered requests for external AI validation.

Faculty Training Tool

Trigger-Response Simulator

Experience the Trigger-Request loop. Act as a student group encountering a challenge and see how GenAIRL responds to support socially shared regulation (SsRL).

Step 1: Identify the Trigger

Your group is stuck. What is the problem?

GenAIRL v2.0 // Collaborative Mode
smart_toy
Hi team! I'm ready to help with your inquiry project. I can review your work, help plan next steps, or clarify concepts. What do you need?

Live Analysis

Trigger Type -
Request Pattern -

The CHARM Loop

event

Trigger

Observable event signals need

arrow_downward
psychology

Recognition

Group identifies regulatory gap

arrow_downward
send

Request

Formulate query to GenAIRL

arrow_downward
smart_toy

AI Act

Detect, Diagnose, Support

arrow_downward
group

SsRL

Interpret & Negotiate support