Scaffolding Metacognition with GenAI: Exploring Design Opportunities to Support Task Management for University Students with ADHD

Paper Title

Scaffolding Metacognition with GenAI: Exploring Design Opportunities to Support Task Management for University Students with ADHD

Publication Info

• Topic area: Leveraging Generative AI (GenAI) to support task management and metacognitive processes for university students with ADHD.
• Keywords: ADHD, task management, metacognition, Generative AI, cognitive scaffolding, emotional regulation, university students, productivity tools, co-design, neurodivergence.

Background and Problem

• Problem / challenge: University students with ADHD face significant challenges in academic task management due to difficulties in metacognition, including limited task awareness, poor self-monitoring, and emotional regulation. Existing GenAI tools are not designed to address these specific needs.
• Significance: Effective task management is critical for academic success, and addressing these challenges can reduce risks of underperformance, dropout, and emotional distress among students with ADHD.
• Motivation and related work: Prior research has explored metacognitive interventions and assistive technologies for ADHD but has not fully examined how GenAI can support metacognitive processes. Current GenAI tools often cater to neurotypical users and may inadvertently undermine metacognition by fostering dependency or reinforcing biases.

Solution

• Proposed approach: Using Generative AI to scaffold metacognitive processes for university students with ADHD, focusing on enhancing task awareness, reflective execution, and emotional regulation.
• Novelty:
  1. Identification of metacognitive challenges specific to ADHD in task management, including fragmented task awareness and emotional barriers.
  2. Co-designed GenAI-based solutions tailored to ADHD-related needs, validated through expert feedback.
  3. Design principles for leveraging GenAI to promote reflection, balance interests with reality, and support emotional growth without fostering dependency.
• Procedure and key techniques:
  1. Conducted co-design sessions with 20 university students with ADHD to explore their task management challenges and envision GenAI solutions.
  2. Interviewed five ADHD experts to evaluate the feasibility, clinical relevance, and potential risks of the proposed designs.
  3. Thematic analysis of participants’ challenges and design ideas, leading to three key design directions for GenAI support.

Results

• Concrete findings:
  • Identified three major metacognitive challenges: lack of task and self-awareness, difficulties in monitoring and control, and emotional barriers.
  • Proposed GenAI solutions include task integration from fragmented sources, personalized time estimation, adaptive prioritization, reflective task execution, and emotional regulation through virtual companions.
• Advantage over baselines:
  • GenAI offers unique capabilities for integrating fragmented task information, providing personalized feedback, and fostering emotional regulation, which traditional tools lack.
  • Experts highlighted the potential for GenAI to reduce cognitive load and promote self-reflection, though they cautioned against risks like dependency and reinforcement of biases.
• Experiments / evaluation:
  • Co-design sessions revealed diverse user needs and design ideas, while expert interviews validated the feasibility and highlighted risks.
  • Participants’ ideas were grounded in real-world academic challenges, and experts emphasized the importance of balancing automation with user agency.
• Limitations and future work:
  • Cultural homogeneity of participants (Mandarin-speaking, Chinese social media users) may limit generalizability.
  • Risks of GenAI dependency and hallucinations need to be addressed in future implementations.
  • Future work will focus on developing a GenAI-based task management tool, incorporating adaptive and privacy-preserving features.

Summary

This study explored how Generative AI (GenAI) can scaffold metacognitive processes for university students with ADHD, addressing task management challenges like fragmented task awareness, poor self-monitoring, and emotional barriers. Through co-design sessions with 20 students and interviews with five ADHD experts, the authors identified three design directions: enhancing task and self-awareness, promoting reflective task execution, and facilitating emotional regulation. While GenAI shows promise in reducing cognitive load and fostering reflection, experts cautioned against risks like dependency and bias reinforcement. The findings provide a roadmap for designing inclusive, metacognition-supportive GenAI tools that balance automation with user agency.