Actionable companion · Self-regulation & metacognition
From AI evidence to responsible practice: Perceived satisfaction, perceived usefulness, and interactive learning environments as predictors of university students’ self-regulation in the context of GenAI-assisted learning: an empirical study in mainland China
AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not findings 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 participants, intervention and comparator before applying this study in a new setting.
Original authors’ abstract, reproduced under the linked source license:
Abstract Given the potential risks of learners’ misuse of generative artificial intelligence (GenAI), including over-reliance, privacy concerns, and exposure to biased outputs, it is essential to investigate university students’ self-regulation in GenAI-assisted learning. Self-regulated learning enables university students to set goals, monitor their learning progress, and adjust strategies, thereby enhancing the effectiveness of GenAI-assisted learning. Guided by the three-tier model of self-regulation, which encompasses individual characteristics, cognitive and emotional factors, and behavioral intention, this study employed a mixed-method approach. Structural equation modeling (SEM) was used to quantitatively examine the relationships among key variables, while interviews provided qualitative insights, enabling a comprehensive exploration of factors influencing self-regulation in GenAI-assisted learning. Using a sample of 607 university students (e.g., prospective mathematics teachers) from Mainland China, this study found that compared to perceived self-efficacy and interactive learning environments, information system quality showed a stronger influence on learners perceived usefulness and satisfaction in GenAI-assisted learning. In predicting learner perceived self-regulation, perceived usefulness was a stronger predictor than the interactive learning environment and perceived satisfaction. Similarly, perceived usefulness was a stronger predictor of behavioral intention than perceived satisfaction and self-regulation. This study further investigated the partial mediating effects of perceived usefulness, perceived satisfaction, and perceived self-regulation among other variables. This study proposes a conceptual model to explore the interconnectedness of these factors in GenAI-assisted learning. It highlights the importance of information system quality for educators and recommends that researchers further investigate the dynamic factors influencing self-regulation in GenAI-assisted learning environments. Keywords: GenAI-assisted learning, perceived self-regulation, structural equation modeling, three-tier models of self-regulation, prospective mathematics teachers
What kind of evidence is this?
Supporting AI learning science · Qualitative / mixed methods
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
This study employed SEM and qualitative interviews to examine the relationships among variables in GenAI-assisted learning. The methodological design was guided by Grant and Osanloo’s (2014) framework, which emphasizes that a theoretical framework serves as the blueprint of a research project. Their guidelines highlight the importance of (a) grounding a study in a clearly defined theoretical model, (b) ensuring alignment between the research questions, hypotheses, literature review, instruments, and data analysis procedures, and (c) providing a clear rationale for the selection of the theoretical foundation. Compared with other general methodological recommendations, such as Creswell and Creswell’s (2017) research design framework or Yin’s (2018) guidelines for case study research, Grant and Osanloo’s (2014) approach places a more explicit emphasis on integrating theory at every stage of the research process. This characteristic aligns with the current study, which requires a strong theoretical grounding to investigate the psychological and contextual factors affecting learners’ self-regulation in GenAI-assisted learning.
Limitations and uncertainty · source passage
From a theoretical perspective, this study offers empirical insights into learners’ self-regulation in GenAI-assisted learning, addressing a gap in existing literature and strengthening the theoretical foundation for applying GenAI in interdisciplinary learning. From a practical perspective, teachers are encouraged to integrate high-quality GenAI tools (e.g., ChatGPT in global contexts and Ernie Bot in China) into their teaching practices to enhance students’ learning experiences. By using these tools, educators can create personalized learning environments where students receive tailored feedback based on their individual needs. This integration enhances learners’ engagement by allowing them to interact with dynamic and creative content (e.g., AI-driven simulations, interactive lessons, and innovative assignments). Second, learners can actively engage in discussions with teachers to optimize their GenAI-assisted learning experiences (e.g., ChatGPT and Ernie Bot), adjusting timely their learning behaviors. Learners should also prioritize identifying valuable information and leveraging the high-quality resources provided by GenAI to develop a deeper understanding. Moreover, students can use GenAI as supportive tools for setting learning goals, managing their learning schedules, and monitoring progress, thereby enhancing their self-regulation. Third, operators should develop emotion-sensing systems to identify learners’ emotional states. Developers should prioritize user experience and system stability by creating learning systems that are user-friendly and accessible. Finally, developers can incorporate dynamic, interactive command designs into ChatGPT (e.g., layered prompts, visual navigation) to guide users step-by-step through operations. This approach not only reduces the learning threshold but also enhances the usability of ChatGPT.
Design the system before the pilot
Use this study to inform a learning mechanism or implementation question. It is not direct evidence that autonomous agents improve learning. Keep the proposed agent behavior separate from what the source actually tested.
- Use plan–attempt–check–reflect prompts and compare confidence with actual performance. Keep learner choice and reflection visible.
- Measure engagement separately from achievement. Check whether novelty-driven participation persists and whether it translates to independent performance.
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
Zhiwei Liu; Yan Zhao; Haode Zuo; Yongjing Lu. Perceived satisfaction, perceived usefulness, and interactive learning environments as predictors of university students’ self-regulation in the context of GenAI-assisted learning: an empirical study in mainland China. Frontiers in Psychology (2025-12-03). https://doi.org/10.3389/fpsyg.2025.1599478
Companion prepared 2026-11-01. 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.