Adaptive Cognitive Scaffolding: Real-Time GenAI Tools That Elicit—and Evolve With—User Metacognition

Building on the CHI 2025 Tools for Thought workshop synthesis—which calls for theory-driven and design-oriented frameworks to both understand and augment human cognition—this idea zooms in on the intersection of generative AI and metacognition, but with a crucial twist: adaptivity. While prior work explores how generative AI tools might prompt critical thinking, offer stepwise support, or provide personalized feedback, these approaches typically assume static intervention strategies or one-size-fits-all guidance. This proposal leverages GenAI’s real-time sensing and language processing to explicitly track markers of metacognition—such as self-questioning, reflection-in-action, or evidence of hypothesis testing—and dynamically adjust its level and type of scaffolding. For example, a GenAI writing assistant could monitor whether the user is weighing alternatives, asking “why” or “how” questions, or explaining decisions, and adapt its prompts accordingly: becoming less directive when users are metacognitively engaged, or increasing prompts for justification and self-explanation when reflection lapses. This approach treats the user's cognitive state as a moving target and positions the AI as an adaptive coach fostering metacognitive resilience even as abilities or fatigue fluctuate. Theoretically, it blends frameworks from metacognitive scaffolding in education, interactive learning analytics, and adaptive AI interaction, while addressing concerns about cognitive atrophy from over-reliance on AI by maintaining human control and reflection. Empirically, it could be tested in domains like writing, coding, or education to evaluate effects on long-term retention, transfer, independent critical thinking, and mitigation of design fixation or passive knowledge absorption. This research could redefine GenAI augmentation by shifting from static aids to dynamic, co-evolving cognitive partners, offering practical design blueprints for tools that cultivate users’ self-awareness, flexibility, and lifelong learning capacity.

References:

1. Understanding, Protecting, and Augmenting Human Cognition with Generative AI: A Synthesis of the CHI 2025 Tools for Thought Workshop. Lev Tankelevitch, Elena L. Glassman, Jessica He, A. Kittur, Mina Lee, Srishti Palani, Advait Sarkar, Gonzalo A. Ramos, Yvonne Rogers, Hari Subramonyam (2025). arXiv.org.