Mastering AI tools and learning skills boosts student writing and well-being

A new peer-reviewed study has revealed that both AI literacy and self-regulated learning (SRL) are critical predictors of university students’ writing performance and psychological well-being in the context of generative artificial intelligence (GAI)-supported education. The study, titled “Exploring How AI Literacy and Self-Regulated Learning Relate to Student Writing Performance and Well-Being in Generative AI-Supported Higher Education”, was published in Behavioral Sciences.

Based on a sample of 257 university students in China, the study used structural equation modeling to assess the direct and indirect relationships between learner traits and outcomes. The findings confirm that students with higher AI literacy and SRL skills achieve better writing performance and experience greater GAI-driven digital well-being, defined as the sustained emotional and cognitive benefits derived from using generative AI tools in academic settings.

How do AI literacy and self-regulated learning impact writing performance?

The study identified both AI literacy and SRL as significant positive predictors of student writing performance. Self-regulated learning, encompassing students’ ability to plan, monitor, and reflect on their academic behavior, showed the strongest effect. Students who engaged in active learning management reported enhanced idea generation, grammatical accuracy, and overall language expression when using GAI tools such as ChatGPT.

AI literacy, measured across four core dimensions (awareness, use, evaluation, and ethics), also contributed significantly to writing outcomes. Students who could critically select, adapt, and combine AI tools for academic purposes consistently reported improved clarity, coherence, and vocabulary usage in their written work.

These findings support the theoretical claim that writing in AI-enhanced environments is not a passive process but one requiring active cognitive and metacognitive engagement. Students with higher SRL were more adept at integrating AI feedback into their work, while those with stronger AI literacy could leverage the tools more effectively for task-specific demands.

What is the role of writing performance in student well-being?

Beyond academic achievement, the study explored the emotional and psychological impacts of GAI use, referred to as GAI-driven well-being. Writing performance was found to serve as a partial mediator between both learner traits (AI literacy and SRL) and well-being outcomes.

Students who performed well in writing tasks using GAI tools reported higher levels of digital well-being. This included a stronger sense of control, satisfaction with the learning process, and reduced academic anxiety. Importantly, AI literacy was shown to have a direct effect on digital well-being (β = 0.503, p < 0.001), while SRL affected well-being indirectly through its impact on performance.

These results reflect the principles of the Control-Value Theory of Achievement Emotions, which posits that well-being is shaped by individuals’ perceived control over academic outcomes. When students feel confident in using AI tools and can regulate their learning effectively, they are more likely to experience positive emotions and engagement during writing tasks.

However, the study cautions that well-being gains are not guaranteed. The data suggest that students with low AI literacy or weak SRL strategies may struggle to derive value from GAI tools, potentially experiencing frustration or disengagement. Thus, while AI tools offer emotional support and cognitive enhancement, their effectiveness depends on the learner’s competencies and mindset.

What are the educational implications of the findings?

The study's integrated model provides strong empirical backing for educational policy shifts that emphasize digital competency and self-regulated learning in AI-rich academic environments. Institutions are encouraged to design curricula that explicitly teach AI literacy, including ethical reasoning, critical evaluation, and creative application of generative AI tools.

Similarly, embedding SRL strategies into instructional design, such as encouraging students to maintain reflective logs, plan GAI-assisted writing projects, and self-assess AI feedback, can promote long-term learning gains. Educators are also urged to move beyond tool-centric training and cultivate a developmental view of writing as a process influenced by both technology and learner agency.

The research further supports the implementation of ethical frameworks and usage guidelines for AI in education. With growing concern over academic integrity, hallucinated AI content, and data privacy, students need structured support to use these tools responsibly. Digital well-being cannot be assumed as a byproduct of access; it must be consciously nurtured through pedagogical and institutional safeguards.

Despite the study’s strengths, the authors acknowledge limitations. The sample was limited to Chinese universities, suggesting the need for cross-cultural replication. Additionally, the use of cross-sectional data means that causality cannot be definitively established. Future studies should pursue longitudinal approaches and include factors like prior academic achievement, GAI access frequency, and socio-economic context.