Actionable companion · Retention & transfer
Learning with agency: AI-Mediated Informal Digital Learning of English for Speaking Development: Longitudinal Effects on Ability and Affect and Implications for Adaptive Intelligence
AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.
Evidence and context
Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Higher education. Verify the participant population and task before transferring the findings.
Original authors’ abstract, reproduced under the linked source license:
Abstract Generative artificial intelligence (GenAI) has expanded opportunities for English learning beyond formal classroom settings. From an intelligence perspective, speaking in English as a foreign language (EFL) represents an applied domain in which learners must adaptively retrieve, integrate, monitor, and deploy linguistic knowledge under communicative demands. The present study examined the effects of a pedagogically guided form of AI-mediated informal digital learning of English (AI-IDLE) on Chinese university EFL learners’ speaking ability, speaking anxiety, and speaking enjoyment. To this end, 89 Chinese university EFL learners participated in a 12-week intervention and were assigned to an experimental group (EG) or a control group (CG). Pre- and post-intervention evaluations were administered to assess the results, including standardized speaking skills and valid questionnaires regarding enjoyment and speaking anxiety. The English-speaking skills test and the two questionnaires were administered at post-test and again 10 weeks after the intervention as the delayed post-test. Mixed-effects modeling (MEM) was used, and the findings showed that the EG demonstrated significantly greater improvements in speaking ability and enjoyment and a greater reduction in speaking anxiety than the CG, with these between-group advantages evident at the post-test. Overall, the findings suggest that structured AI-IDLE can provide a digitally mediated context for intelligence-relevant adaptive learning by supporting communicative performance while fostering affective conditions conducive to its development. These findings have implications for understanding how digitally mediated informal learning environments may support adaptive intelligence in applied language-learning contexts. Keywords: AI-mediated informal digital learning of English (AI-IDLE), adaptive intelligence, speaking ability, speaking anxiety, speaking enjoyment
What kind of evidence is this?
Intelligent tutoring & adaptive learning · Comparative / pre-post study reported
Stage: Higher education. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.
Study methods · source passage
Participants within the present research included 89 EFL students in China who registered in English courses in a Chinese college. Convenience sampling was used to select the learners because they were available to the scholars and were in the same speaking-oriented English course within the semester where the research was performed. The participants ranged in age from 18 to 22 years and represented different academic majors. Although some participants reported occasional personal use of AI tools, they had limited experience using GenAI as part of a structured classroom learning environment. This characteristic helped establish a relatively comparable starting point for examining the instructional effects of the intervention. The study adopted a quasi-experimental longitudinal design. Participants were recruited within the same speaking-oriented course context rather than from multiple independently sampled instructional settings. Following convenience recruitment, the 89 participants were allocated to an experimental group (EG; n = 45) or a control group (CG; n = 44). Because individual random assignment was not employed, baseline equivalence between the two groups was examined statistically before the intervention. The former received speaking instruction through GenAI tools, while the latter followed conventional speaking activities based on the regular curriculum. Independent-samples comparisons at pretest indicated no statistically significant group differences in speaking ability, speaking anxiety, or speaking enjoyment, suggesting comparable baseline levels on the study outcomes.
Limitations and uncertainty · source passage
The present results suggest some directions for subsequent research. Basically, longitudinal research is needed to track whether reducing speaking anxiety and increasing speaking enjoyment, as well as the capability offered by GenAI, can be maintained over extended teaching periods. Moreover, the current research concentrated on voice mode, though future research can study the multimodal incorporation of GenAI with other tools, like video, to investigate the way enhanced immersion influences the fulfillment of psychological needs. Subsequent research needs to take a comparative method to investigate various student profiles to specify the way personal differences in enjoyment or anxiety interplay with GenAI efficacy. Finally, research needs to concentrate on the repetitive co-design of AI tools, guaranteeing that the development of GenAI characteristics is informed by the particular emotional and educational needs of EFL students. These efforts can guarantee that GenAI acts as a strong support for psychological concepts and linguistic development in language teaching.
Design the system before the pilot
Document the learner model, mastery rule, hint sequence and adaptation signal. Compare adaptive routing with the same content in a non-adaptive sequence. Check whether early errors lead to persistent low-expectation pathways.
- Include a delayed assessment and an unfamiliar transfer problem without AI. Immediate supported performance cannot establish durable learning.
- Use common rubrics, blind scoring where feasible and a fresh unaided reasoning task. Document permissible assistance and avoid treating AI-detector scores as proof of misconduct.
A 90-day district pilot
- Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
- Days 16–30 · Design. Choose one disciplinary task. Require a first attempt, a documented AI consultation, a critique of the output and an independent explanation. Align the comparison group and scoring rubric with the same objective. Name an accountable owner and set the baseline, success threshold and budget before implementation.
- Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
- Days 61–90 · Decide. Blind-score reasoning and accuracy on a new task without AI. Distinguish model examination scores, student perceptions and actual learner gains; only the last directly tests educational benefit. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.
The board decision brief
Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.
Attribution and reuse
Difei Jia; Xi Chen; Yunsong Wang. AI-Mediated Informal Digital Learning of English for Speaking Development: Longitudinal Effects on Ability and Affect and Implications for Adaptive Intelligence. Journal of Intelligence (2026-09-01). https://doi.org/10.3390/jintelligence14090204
Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.