
Abstract Lexical errors are among the most frequent and persistent issues in second language (L2) production, yet research on vocabulary exercises has largely focused on the acquisition of new words rather than on the correction of misused ones. This exploratory repeated-measures study investigates the potential of generative artificial intelligence (GenAI) to provide personalized vocabulary practice targeting learner-specific lexical errors in Spanish. Eighty-five L2 and heritage learners across proficiency levels completed 15 individualized exercises generated by a custom GPT-based system from their own writing errors. The tasks varied in goal, format, linguistic complexity, and involvement load (IL). Error recognition and correction were assessed at immediate and delayed post-tests, and mixed-effects models showed that higher exercise performance and task-induced IL predicted better outcomes. Both recognition and correction declined at the delayed post-test, although only recognition did so significantly. Among exercise features, productive, contextualized formats such as text editing, translation, and sentence reconstruction showed strong positive associations with correction, while goals emphasizing paraphrastic flexibility were less favorable. At the learner level, outcomes also varied substantially. Overall, the findings suggest that GenAI can support scalable, individualized, error-driven vocabulary practice. Still, variability across learners and exercises, together with incomplete retention, indicates that personalization alone may not ensure durable learning. Future research should thus examine whether feedback, recycling, and adaptive review improve retention.
Abstract Retrieval-based digital flashcards have proved their effectiveness in L2 vocabulary learning. However, recent developments in AI technology have introduced AI-assisted vocabulary learning apps that integrate multimedia input and AI-supported output activities. Considering these technological advances, this study compared the effects of an AI-assisted multimedia app (WordUp) and a retrieval-based flashcard app (Quizlet) on university students’ receptive vocabulary. A total of 128 students from three EFL classes were assigned to two experimental groups and one control group. Each experimental group used its assigned app to study target words. The control group did not receive any intervention. Students’ learning gains were measured with a vocabulary test administered as an immediate and a delayed posttest. Semi-structured interviews with 20 participants provided insights into learners’ experiences. The findings indicated that both WordUp and Quizlet notably enhanced receptive vocabulary knowledge. However, the WordUp group demonstrated significantly higher vocabulary learning gains than Quizlet in the immediate and delayed posttests. The control group showed negligible improvement. Interviews revealed that learners valued Quizlet’s simple interface and structured review but also expressed a need for greater interactivity and multimedia support. In contrast, WordUp was valued for its contextualized multimedia resources and AI-assisted productive activities, though learners reported challenges related to interface and content complexity. Overall, the findings suggest that effective mobile-assisted vocabulary learning depends not only on the technological features available in learning apps but also on balancing rich multimedia and interactive learning opportunities with usability and learners’ cognitive readiness.
Abstract Research on informal digital learning of English (IDLE) has grown substantially, yet a systematic analysis of its methodological foundations is lacking. To map current practices, this methodological review examines 104 peer-reviewed studies (2015–2025) indexed in the Arts & Humanities Citation Index or the Social Sciences Citation Index . The findings reveal a dominant pattern of methodological homogeneity , characterized by heavy reliance on non-experimental designs (94.23%), self-report instruments (notably questionnaires; 69.49% of quantitative tools), and descriptive and correlational analyses (80.44% of quantitative analysis strategies). This profile is understandable, as the informal, nonlinear, and learner-driven nature of IDLE aligns naturally with exploratory approaches. However, this focus may limit the investigation of causal mechanisms and theoretical development. To support the field’s evolution, we propose a strategic diversification of methods. Integrating longitudinal designs to trace development, complementing self-reports with ethically collected and activity-sensitive behavioral and trace data, and leveraging new analytical tools such as learning analytics and, where appropriate, AI-assisted analysis can enrich inquiry and help build more robust, explanatory models of informal digital learning.
Abstract Recent advances in generative artificial intelligence (GAI) have attracted increasing attention as an alternative interlocutor to promote willingness to communicate (WTC) in a second language (L2). However, learners’ ability to benefit from AI-mediated interaction partly depends on their capacity to formulate effective prompts, while systematic instruction in prompting strategies remains limited. Rather than replacing peer communication, GAI may function as a preparatory resource that supports learners in building confidence before engaging with classmates. The present study investigates prompt-guided human–GAI–human learning, an instructional approach that integrates GAI interaction into communicative learning cycles. Specifically, the study examines whether prompt instruction, compared with self-directed GAI use, enhances learners’ WTC with voice-enabled GAI during task preparation and with peers during subsequent discussions. Eighty first-year university EFL students were assigned to either a prompt-instruction condition or a self-directed GAI-use condition. Over six weeks, both groups completed weekly oral discussion tasks following a human–GAI–human sequence. WTC questionnaires were administered at three time points, and reflective accounts were collected to examine learners’ perceived influences on communication. Both groups showed increases in WTC over time, while prompt instruction was associated with higher overall WTC with GAI. However, prompt instruction did not significantly alter the trajectory of WTC development. The overall increase in WTC with peers across both groups suggests that AI-supported preparation may facilitate learners’ communicative readiness for subsequent peer interaction. These findings highlight the pedagogical value of prompt-guided human–GAI–human learning for supporting L2 communication.
This study explores app development as an instructional medium for fostering perceived digital competence and pedagogical authorship in English teacher education, contributing to research on technology-enhanced preparation that moves teachers from using digital tools towards creating them. Participants developed simple instructional apps through low-threshold coding activities in iterative group design cycles across two groups - undergraduates in-person (n = 18) and graduates online (n = 16) - sharing a 15-week curriculum. Pre-post surveys indicated significant gains in digital confidence and perceived instructional benefit among undergraduates, while graduates showed clearer movement in perceived instructional benefit, with engagement and motivation remaining stable across both groups. Lexical cluster analysis of reflection essays identified four orientations - emotional growth, pedagogical insight, design conceptualization, and scaffolding reliance - pointing towards emerging pedagogical authorship rather than uniform competence gains. As digital creation becomes as common as digital consumption for today's learners, modest coding literacy, framed as an instructional medium, offers a workable path towards digital agency and pedagogical authorship in English teacher education.