A Framework for Generative AI-based Cognitive Engagement

• In Event: Innovating K–12 Teaching and Learning with Generative AI: Insights From IES-funded National R&D Centers

Abstract

Objective
Cognitive engagement involves the sustained mental effort students apply to understand and use knowledge, shaping curiosity, motivation, and knowledge transfer (Chi & Wylie, 2014). Teachers face challenges tailoring instruction to diverse learner needs, including prior knowledge and experiences (Kang et al., 2005). Generative AI (GenAI) offers a powerful solution by analyzing student profiles in real time and generating adaptive, multimodal content (Xie et al., 2022). This study proposes a GenAI-based framework that integrates four strategies—cognitive conflict, analogy, experience bridging, and engaged critiquing—with GenAI’s capabilities, enabling the creation of responsive, engaging learning materials that support engagement.

Perspectives
Cognitive conflict engages students by challenging misconceptions with contradictory information, prompting conceptual change through reflection and inquiry (Chinn & Brewer, 1993); Analogical reasoning aids learning by mapping structural similarities between familiar and unfamiliar concepts, using a framework of entities, configurations, and mechanisms (Rivet & Kastens, 2012); Effective analogies require analyzing concept complexity, evaluating prior knowledge, and ensuring familiarity with the analogue (Treagust et al., 1998); Experience bridging connects abstract ideas to students’ real-life contexts, enhancing relevance and engagement (Lundholm et al., 2013): and Engaged critiquing promotes critical thinking through evidence-based evaluation, enabling students to refine reasoning and engage in scientific discourse.

GenAI has the capacity to generate learning materials adapting to cognitive profiles, behaviors, and cultural contexts (Liang et al., 2023; Owoseni et al., 2024), which deepen understanding and sustain engagement (Anyichie et al., 2023). GenAI can generate multimodal materials like text, images, simulations, and videos, aligning with students’ preferences (Lee et al., 2025).

GenAI-based Cognitive Engagement Framework

Assessing students’ pre-knowledge
This phase involves GenAI to assess students’ pre-knowledge such as prior knowledge, misconceptions, and learning behaviors on scientific concepts, aligning with curriculum standards and instructional objectives.

Determine cognitive engagement strategy
GenAI determines the optimal strategy by integrating students’ pre-knowledge, background, and the nature of disciplinary ideas. Table 1 summarizes how each factor influences the strategy choice. Four strategies are emphasized:
Cognitive Conflict is effective when misconceptions are common. GenAI creates scenarios or counterexamples to challenge erroneous beliefs.
Analogical Reasoning maps unfamiliar concepts onto familiar ones. GenAI constructs analogies based on structural similarity and learner background.
Experience Bridging connects abstract ideas to students’ lived experiences. GenAI uses cultural and contextual data to localize content.
Engaged Critiquing promotes evaluation and refinement of ideas through argumentation. GenAI scaffolds this by assessing readiness and generating critique-based tasks.

Determining modalities
GenAI adapts instructional content using text, images, audio, video, or interactive tools based on learners’ profiles and. For example, it may create analogical illustrations for visual learners or simulations for kinesthetic learners, ensuring alignment with both the learner and the engagement strategy.

Generating learning materials
Synthesizing multimodal materials and adopting to students’ cognition, GenAI generates engaging learning materials to promote student engagement.

Significance
This study introduces a framework for using GenAI to enhance cognitive engagement in K–12 classrooms. By leveraging GenAI’s personalization, multimodality, and adaptive scaffolding, the framework empowers teachers to design responsive and inclusive instruction. It underscores the need for ethical implementation and ongoing educator-AI collaboration to maximize learning outcomes.

Authors

• Lehong Shi, University of Georgia

• Shuchen Guo, Nanjing Normal University

• Yizhu Gao, University of Georgia

• Hongmei Li, Yunnan Normal University

• Liang Zhang, University of Georgia

• Ninghao Liu, University of Georgia

• Xiaoming Zhai, University of Georgia