Exploring the Co-Intelligence Gap in ChatGPT Utilization between Developed and Low and Middle Income Countries: An Experimental Design Based on Mollick’s Four Principles of AI Collaboration

Generative AI is evolving beyond a simple tool into a “Co-Intelligence” partner that collaborates with humans, serving as a fundamental basis for the progression toward Artificial General Intelligence (AGI). However, this advancement implies the potential to restructure inequalities based on differences in AI utilization capabilities. In particular, disparities in national educational environments and digital resources create unequal opportunities to utilize AI Co-Intelligence, which may lead to future knowledge and technological gaps. This study explores the existence and implications of a “Co-Intelligence Gap” by comparing the usage patterns, strategies, and learning outcomes of learners from developed and developing countries who performed identical tasks using ChatGPT. Over a one-month period, the study surveyed participants’ perceptions of integration ChatGPT into their learning processes and analyzed learning outcomes, AI trust, self-efficacy, usage behaviors based on self-reported measures and perceptions of collaboration principles. The analysis identified patterns in usage depth, strategies, and outcome differences between the two groups while controlling for confounding variables such as English proficiency and digital literacy. This study is significant for empirically examining the structural differences in AI collaboration capabilities across nations and expanding the digital divide discourse from accessibility to a capability-centered paradigm, thereby providing foundational data for future Co-Intelligence-based pedagogical models.