Abstract
Artificial intelligence (AI) tools, such as ChatGPT, are becoming increasingly embedded in higher education for assisting students in tasks ranging from information retrieval to essay writing and problem solving in China. Although such tools provide powerful cognitive support, the growing dependence on them raises concerns on the types of cognitive tasks students choose to offload to AI tools and the conditions under which such offloading occurs. Drawing on a value-based decision-making framework, this work investigated the factors influencing AI-based cognitive offloading by university students. Two sequential studies were conducted. Study 1 developed and validated a scale that distinguishes between AI-based lower- and higher-order cognitive offloading. Three-sample analyses indicated that the scale exhibited robust psychometric properties. Study 2 tested the relationship between the perceived value of AI, AI-usage capability, AI evaluation capability, and both types of AI-based cognitive offloading by analyzing the behavior of 667 university students. All data were collected in mainland China. Results indicate that the perceived value of AI is positively associated with both types of cognitive offloading through AI-usage capability. This indirect association was conditioned by the AI evaluation capability: it became nonsignificant for lower-order offloading under a high evaluation capability but remained significant for higher-order offloading. These findings highlight the conditional nature of AI-based cognitive offloading in higher education. Moreover, they extend the cognitive offloading theory from tool-based to human–AI collaborative contexts, offering theoretical and practical insights for fostering reflective and value-driven human–AI interaction in learning environments.
Showing the abstract — retrieve the full paper via the Exa API.