Can DeepSeek Intelligent Chatbot’s Emotion Induction Enhance College Students’ English Learning Ability?

Chatbots have become increasingly popular as tools for supporting English language learning. However, how to effectively enhance learning ability through emotion induction during interactions with intelligent chatbots remains an unresolved issue. This study integrates Flow Theory, Perceived Value Theory, and the Unified Theory of Acceptance and Use of Technology (UTAUT) with emotion-induction features of a chatbot to develop and test a theoretical model. Survey data from 388 university students who used DeepSeek for English learning were analyzed using structural equation modeling (SEM). Learning ability was operationally defined as continuance learning behavior and learning engagement or self-rated performance, measured through Likert-scale items. The results show that performance expectancy, effort expectancy, and perceived emotional value were all significantly associated with flow, with effort expectancy exhibiting the most substantial direct effect on flow. Perceived emotional value exerts the most significant total effect on continuance learning behavior by enhancing emotional engagement and reinforcing sustained effort through flow. Positive affective prompts strengthen effects of performance expectancy, effort expectancy, and perceived emotional value on flow. In contrast, negative emotion induction appears to reduce the strength of these associations. These findings specify which cognition–affect pathways matter most for English learning with chatbots and inform the design of emotion-aware prompting in DeepSeek to enhance flow and continuance learning behavior.