Modeling AI‐Assisted English Academic Writing: Insights From L2 Motivational Self System and Technology Acceptance Model

ABSTRACT This study examined the influence of second/foreign language (L2) writing selves, English academic writing learning experience (WLE), and education level (EDUL) on AI adoption by English‐as‐a‐foreign‐language learners in English academic writing. Perceived usefulness (PU), perceived ease of use (PEU), attitude towards use (ATU), behavioral intention (BI), the ideal and ought‐to L2 writing selves, and WLE were selected as latent variables. Hypotheses were established based on a review of relevant studies. Quantitative data were collected via an online questionnaire from 316 participants (166 undergraduates and 150 postgraduates) and analyzed using structural equation modeling. Our findings revealed that although L2 writing selves positively shaped their perceptions of AI, neither the ideal/ought‐to selves nor WLE directly predicted their intentions to use such tools. While the ideal self indirectly influenced ATU and BI through PEU and PU, the indirect effect of the ought‐to self was mediated by PU. Additionally, EDUL was found to positively moderate two pathways. The model explained 76.1% of the variance in BI. Through integrating the L2 motivational self system (L2MSS) and the technology acceptance model (TAM) in the English academic writing context, this study provided pedagogical suggestions to English academic writing teachers.