Next Article in Journal
Non-Semantic Multimodal Fusion for Predicting Segment Access Frequency in Lecture Archives
Previous Article in Journal
Harnessing Intelligent GISs for Educational Innovation: A Bibliometric Analysis of Real-Time Data Models
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Artificial Intelligence-Generated Content Empowers College Students’ Critical Thinking Skills: What, How, and Why

School of Education, Hunan University of Science & Technology, Xiangtan 411201, China
*
Author to whom correspondence should be addressed.
Educ. Sci. 2025, 15(8), 977; https://doi.org/10.3390/educsci15080977
Submission received: 4 July 2025 / Revised: 25 July 2025 / Accepted: 28 July 2025 / Published: 30 July 2025

Abstract

Despite the increasing number of studies indicating that generative artificial intelligence is conducive to cultivating college students’ critical thinking skills, research on the impact of college students’ use of generative artificial intelligence on their critical thinking skills in an open learning environment is still scarce. This study aims to investigate whether the use of generative artificial intelligence by college students in an open learning environment can effectively enhance their critical thinking skills. The study is centered around the following questions: Does the use of generative artificial intelligence in an open learning environment enhance college students’ critical thinking skills (what)? What is the mechanism by which the use of generative artificial intelligence affects college students’ critical thinking (how)? From the perspective of self-regulated learning theory and learning motivation theory, what are the reasons for the impact of generative artificial intelligence on college students’ critical thinking skills (why)? To this end, the study employs questionnaires and interviews to collect data. The questionnaire data are subjected to descriptive statistical analysis, correlation analysis, multiple stepwise regression analysis, and mediation effect analysis. Based on the analysis of interview materials and survey questionnaire data, the study reveals the impacts and mechanisms of college students’ use of generative artificial intelligence tools on their critical thinking skills. The findings of the study are as follows. First, the frequency of artificial intelligence use is unrelated to critical thinking skills, but using it for reflective thinking helps to develop critical thinking skills. Second, students with strong self-regulated learning skills are more likely to use generative artificial intelligence for reflective thinking and achieve better development in critical thinking skills. Third, students with strong intrinsic learning motivation are more likely to use generative artificial intelligence for reflective thinking and achieve better development in critical thinking skills. Consequently, the article analyzes the reasons from the perspectives of self-regulated learning theory and learning motivation theory and offers insights into how to properly use generative artificial intelligence to promote the development of critical thinking skills from the perspectives of higher education institutions, college teachers, and college students.

1. Introduction

The application of artificial intelligence-generated content (AIGC), represented by ChatGPT and DeepSeek, is becoming increasingly popular, posing unprecedented challenges to higher education. Artificial intelligence has already had a substitution effect on routine and repetitive physical and mental labor positions. College students trained under the traditional higher education model are facing great challenges in employment. University education needs to shift from homogenization to personalization, paying attention to individual differences among students and cultivating talent with high-order thinking skills such as critical thinking skills. The application of artificial intelligence technology in education is expanding rapidly, and college students with smartphones have become frequent users of AIGC (Okonkwo & Ade-Ibijola, 2021). The existing literature on the impact of AIGC use on critical thinking skills mainly includes two types.
The first type focuses on examining whether the use of AIGC can promote the development of high-order thinking skills, including critical thinking skills, in a strictly controlled classroom environment. These studies usually adopt a quasi-experimental approach, dividing students into control and experimental groups. Students in the experimental group use AIGC as an auxiliary learning tool in the classroom learning process. The results generally support the idea that the use of AIGC can effectively promote the development of high-order thinking skills. For example, Essel et al. (2024) divided 125 students into experimental and control groups. The experimental group integrated ChatGPT as an interactive tool into pre-class exploration, classroom interaction, and task evaluation. The control group did not use ChatGPT, relying on traditional books and the Internet to obtain information. The study found that the use of ChatGPT significantly improved students’ critical, reflective, and creative thinking skills. Other quasi-experimental studies include those by Xu et al. (2024) and Hooshyar and Druzdzel (2024), both of which argue that integrating AIGC into teaching can effectively enhance high-order thinking skills, including critical thinking skills.
The second type explores how the use of generative artificial intelligence tools affects high-order thinking skills by analyzing survey data. These studies reveal the relationship between the use of AIGC or its dimensions and high-order thinking skills, including critical thinking skills, and find that the use of AIGC tools can play a positive role in promoting high-order thinking skills. For example, Li et al. (2025) found, through the analysis of survey data, that an increase in the frequency of use of AIGC tools can enhance high-order thinking skills, including critical thinking skills, and that the impact of deep creative applications is greater. The interaction quality mediates the effect of frequency of use on high-order thinking skills. Similarly, Fadillah et al. (2024) analyzed survey data based on the three core dimensions of ChatGPT (convenience and quality, motivation and engagement, accuracy and trust) and found that ChatGPT can effectively support the development of high-order thinking skills, including critical thinking skills, in physics inquiry by optimizing convenience, enhancing motivation, and ensuring accuracy.
College students’ use of AIGC is mostly a spontaneous behavior, and most students use it in an open learning environment for self-directed learning activities such as completing assignments and writing papers (Yu, 2023). Mogavi et al. (2024) argue that self-directed learning, without external constraints, relies entirely on the learner’s internal motivation and self-discipline and is prone to problems such as overreliance on technology, poor learning habits, and weakened social and critical thinking skills. From the analysis of the above research, it can be seen that there is little consideration in the academic community of how to avoid college students’ overreliance on technology while effectively improving their critical thinking skills in an open learning environment. The purposes of this study are to reveal whether AIGC enhances college students’ critical thinking skills in an open learning environment (what) through the analysis of survey data and interview materials, to reveal its mechanism and path (how), to analyze the reasons from the perspectives of self-regulated learning theory and learning motivation theory (why), and to offer suggestions on how college students can use AIGC reasonably in an open environment.

2. Literature Review and Research Hypotheses

2.1. The Use of AIGC Tools and the Development of Critical Thinking Skills

Critical thinking is a “disposition and skill for reflective skepticism” (McPeck, 1981), comprising two dimensions: cognitive (i.e., skills) and affective (i.e., dispositions) (Ku & Ho, 2010). The cognitive dimension emphasizes logical reasoning and analysis, focusing on an individual’s ability to understand problems and propose rational solutions to the identified issues. This includes the ability to reason, identify assumptions, infer, interpret, analyze, and evaluate arguments (Halpern, 1998). The affective dimension refers to the tendency to engage in an activity or the way that a person approaches a task (Ku & Ho, 2010). Some also argue that critical thinking skills can exist without dispositions, but having dispositions implies the possession of relevant skills (Facione et al., 1995).
Regarding the impact of AIGC on learners, Mogavi et al. (2024) believes that the use of artificial intelligence can lead to overreliance on technology, increase learner laziness, foster superficial learning habits, and reduce the ability for independent thinking (Zhang et al., 2024), thereby weakening their critical thinking skills. However, Zhou et al. (2024) argues that, when AIGC tools are effectively integrated into the curriculum teaching process, they can play a positive role in expanding thinking and support the development of students’ critical thinking and problem-solving abilities. Li et al. (2025) share the same view. They found that only by combining the use of technology with reflective dialog can the development of higher-order thinking skills, such as critical thinking skills, be effectively fostered. In light of the definition of critical thinking skills, reflective thinking plays a key role in the development of critical thinking skills. In other words, when college students apply AIGC to reflective thinking activities, it is possible to promote the development of their critical thinking skills, and there is a strong correlation between the two. Thus, the following research hypothesis is proposed: the application of AIGC in reflective thinking can lead to the development of critical thinking skills. This hypothesis can be verified by analyzing the correlation between the application of AIGC in reflective thinking activities and critical thinking skills.

2.2. The Development of Self-Regulated Learning Abilities and Critical Thinking Skills

Self-regulated learning (SRL) is a process in which learners actively and consciously plan, monitor, regulate, and evaluate their own learning processes in order to achieve learning goals (Panadero, 2017). Students with good self-regulated learning skills can adjust their learning goals and strategies based on the results of self-evaluation, optimize their learning behaviors, and continuously make progress in their studies. Self-regulated learning theory has been proven to be effective in understanding college students’ academic performance and abilities (Cassidy, 2011), and it has been widely verified in both traditional learning environments and computer-supported learning environments.
College students’ use of AIGC occurs more often in an open environment and is a spontaneous behavior. Compared with the classroom teaching environment, the open environment places higher demands on learners’ self-regulation abilities, and students with higher self-regulated learning skills are more likely to improve their critical thinking skills. Zhu et al. (2025) believe that, although AIGC has significant advantages in assisting self-directed learning in open learning environments, its effectiveness depends on students’ prior knowledge and their cognitive and metacognitive levels. The limitations and breakthroughs of AIGC lie in the learners themselves. Multiple studies have found a correlation between self-regulated learning skills and critical thinking skills, with students who have stronger self-regulated learning skills performing better in critical thinking (Bayuningsih et al., 2017). Self-regulated learning includes three key elements: prior domain knowledge, self-efficacy, and the use of learning strategies. Prior domain knowledge provides the basis and material for critical thinking. Self-efficacy gives students more confidence to engage in critical thinking and problem solving. The rational use of learning strategies helps students to analyze, evaluate, and reason about information more efficiently, thereby promoting the development of critical thinking (Widana, 2022). Thus, Hypothesis 2 is proposed: college students with stronger self-regulated learning skills are more likely to use AIGC for reflective thinking. If Hypothesis 1 holds true, then they are more likely to achieve the development of critical thinking skills than students with weaker self-regulated learning skills. Self-regulated learning skills influence critical thinking skills via the application of AIGC to reflective thinking, which means that self-regulated learning skills may have an indirect impact on critical thinking skills through the application of AIGC for reflection. There may be a mediating effect among the three.

2.3. The Development of Intrinsic Learning Motivation and Critical Thinking Skills

Intrinsic learning motivation has a positive predictive effect on the cultivation of critical thinking skills. Students with higher intrinsic learning motivation are more likely to employ deep processing methods when dealing with learning information. This means that they focus more on an in-depth and meticulous analysis and understanding of the information, emphasizing the exploration of its intrinsic meanings, structures, and connections with existing knowledge. In contrast, students with lower intrinsic learning motivation tend to adopt surface processing methods, which means that they concentrate on the superficial features of the information to meet short-term memory or task requirements, without developing an in-depth understanding or analysis of the information (Lu et al., 2021). There is a significant difference in learning outcomes between these two information processing methods. Surface learning methods are usually more associated with habitual behaviors, while deep learning methods can effectively promote the development of higher-order thinking skills such as comprehension and critical thinking skills (Phan, 2006). It can be inferred that students with higher intrinsic learning motivation will employ deep processing methods when using AIGC to delve into its intrinsic meanings and think about the connections with the knowledge that they have already mastered, thereby promoting the development of critical thinking skills. However, students with lower intrinsic learning motivation are likely to still use surface processing methods when using AIGC, being content with quickly obtaining information through AIGC to complete short-term tasks or meet memory needs. Thus, Hypothesis 3 is proposed: students with stronger intrinsic learning motivation are more likely to use AIGC for reflective thinking. If Hypothesis 1 holds true, then they are more likely to achieve the development of critical thinking skills than students with weaker intrinsic learning motivation. Intrinsic learning motivation influences critical thinking skills via the application of AIGC to reflective thinking, which means that intrinsic learning motivation may have an indirect impact on critical thinking skills through the application of AIGC for reflection. There may be a mediating effect among the three.

3. Materials and Methods

3.1. Data Source

This study investigated college students through questionnaires and interviews. The research team conducted the survey at a local university in Central Hunan Province, distributing a total of 628 questionnaires and receiving 566 valid responses, with an effective response rate of 90.13%. The survey questionnaire was self-compiled, with the main items including basic information, AIGC usage, and a critical thinking disposition scale, learning motivation scale, self-regulated learning scale, AI dependency scale, etc. Basic information included gender, grade, major, family economic status, and so on. The questionnaire employed a single-factor test to assess common method bias. The researcher subjected the data from the key items and scale items to exploratory factor analysis (EFA) to determine whether a single factor dominated all items. The principal component analysis revealed that the first unrotated factor explained 23.03% of the variance, which was below the 40% threshold, indicating that the data were not severely affected by common method bias.
The interview outline consisted of 11 questions, divided into five categories. The first category was about the use of AIGC, aiming to understand the basic situation and usage methods regarding AIGC among the interviewees, including four questions: “What AIGC tools do you usually use?”, “How frequently do you use AIGC tools?”, “What are the purposes of using AIGC tools?”, and “Do you reflect on the answers provided by AIGC or use these answers in your reflective thinking?” The second category was about the impact of AIGC usage on critical thinking skills, including two questions: “Do you think the use of AIGC has helped your critical thinking skills?” and “In what situations does this help occur?” The third category was about self-regulated learning abilities, including two questions: “How do you think your learning regulation ability is?” and “Has your learning regulation ability affected your use of AIGC?” The fourth category was about learning motivation, including two questions: “What do you think your learning is for? Is it to be better than your classmates, to meet your parents’ expectations, to compete for better employment opportunities, or to improve your own abilities?” and “Has the above learning purpose affected your use of AIGC?” The fifth category was an open-ended question, asking the interviewees if they had any other thoughts. The research team randomly interviewed five college students and recorded the interview materials in written form.

3.2. Research Variables

3.2.1. Main Variables

Based on the hypotheses put forward earlier, this study takes variables such as the usage frequency of AIGC, the use of AIGC for reflection, self-regulated learning abilities, and intrinsic learning motivation as independent variables. The usage frequency of AIGC was measured by a single multiple-choice question: “How often do you use AIGC?” The options included less than 7 times per week, once a day, about 2–5 times a day, and more than 5 times a day. The use of AIGC for reflection was also measured by a single multiple-choice question: “I reflect on the answers provided by AI in my mind or use them in my reflective thinking.” The options were strongly agree, agree, neutral, disagree, and strongly disagree. Self-regulated learning ability was measured using a self-regulated learning scale adapted from Barnard et al. (2009), which included six dimensions: goal setting, environmental structuring, task strategies, time management, seeking assistance, and self-evaluation. The number of items for each dimension was 5, 4, 4, 3, 4, and 4, respectively, for a total of 24 items. For more details, please refer to Appendix A. The researcher conducted a principal component analysis of the scale using orthogonal varimax rotation in exploratory factor analysis. The results showed that the KMO was 0.928, and Bartlett’s test of sphericity was significant (p < 0.001), indicating that the questionnaire items were well designed and the scale had good validity. The Cronbach’s α coefficient of the self-regulated learning scale was 0.922, indicating good reliability. Intrinsic learning motivation was measured using an intrinsic learning motivation scale adapted from Wang and Chen (2010), which consisted of 3 items. For more details, please refer to Appendix A. In this study, the KMO of the intrinsic learning motivation scale was 0.716, and Bartlett’s test of sphericity was significant (p < 0.001). The Cronbach’s α coefficient of the intrinsic learning motivation scale was 0.756, indicating good reliability. The dependent variable was critical thinking disposition, which was measured using a critical thinking disposition scale adapted from Gu et al. (2025), consisting of 5 items. For more details, please refer to Appendix A. In this study, the KMO of this scale was 0.799, and Bartlett’s test of sphericity was significant (p < 0.001). The Cronbach’s α coefficient of the critical thinking disposition scale was 0.757, indicating good reliability. The above scales used a five-point Likert scale, where 1–5 represented strongly disagree, disagree, neutral, agree, and strongly agree, respectively.

3.2.2. Control Variables

This study includes gender, grade, family economic status, AI dependency, and AI literacy as control variables.
There may be differences in the amount of time that college students of different grades spend using AI, and college students of different majors may have differences in information literacy (Chi et al., 2021). This may lead to different impacts on the development of critical thinking skills. It is generally believed that there is a correlation between critical thinking skills and family economic status. Students from wealthier families have more opportunities to participate in various cultural and academic activities. Their parents may have more time and energy to invest in their children’s education; they are are more likely to adopt scientific educational methods and focus on cultivating their children’s higher-order thinking skills, such as independent thinking, creativity, and critical thinking (Menezes Cunha et al., 2021; Shi & Shen, 2007).
AI dependency refers to an excessive reliance on AI technology and applications in academic research, daily life, and social interactions, manifested as relying on AI to obtain answers when tasks can be completed independently. Long-term AI dependency can lead to increased laziness, reduced creativity, and decreased critical and independent thinking among students (Zhang et al., 2024). AI literacy is a set of skills that enables individuals to critically evaluate artificial intelligence technology, communicate and collaborate effectively with artificial intelligence, and use artificial intelligence as a tool in online, home, and workplace settings (Lu et al., 2025). Research has shown that AI literacy has a direct impact on critical thinking skills and can also have an indirect impact through behavioral engagement and peer interaction (Lu et al., 2025). The AI dependency scale in this study was adapted from Andreassen et al.’s (2012) Facebook dependency scale, with a total of 6 items. In this study, the KMO of this scale was 0.818, and Bartlett’s test of sphericity was significant (p < 0.001). The Cronbach’s α coefficient of this scale in this study was 0.791, indicating good reliability. For AI literacy, Lu et al.’s (2025) AI literacy scale was used, which included four dimensions: awareness, use, evaluation, and ethical guidelines, with a total of 12 items. In this study, the KMO of this scale was 0.872, and Bartlett’s test of sphericity was significant (p < 0.001). The Cronbach’s α coefficient of this scale in this study was 0.845, indicating good reliability. For the specific items of the AI dependency scale and the AI literacy scale, please refer to Appendix A.
The gender variable is represented by 1 for male and 0 for female; grade is represented by 1–5 for freshman, sophomore, junior, senior, and graduate student, respectively; family economic status is represented by 1–5 for very poor, poor, average, wealthy, and very wealthy, respectively; and the AI dependency and AI literacy scales use a five-point Likert scale, with 1–5 representing strongly disagree, disagree, neutral, agree, and strongly agree, respectively.

3.3. Data Analysis Methods

Based on Hypothesis 1, there may be a correlation between the two variables of AIGC applied to reflective thinking and critical thinking disposition. Therefore, after using descriptive statistical analysis to identify the basic information of the sample, this study employs correlation analysis and multiple stepwise regression analysis to examine the survey data. Correlation analysis can reveal whether there is a relationship between two variables, while multiple stepwise regression analysis, by incorporating more relevant variables, can more accurately determine the correlation between the two variables of AIGC applied to reflective thinking and critical thinking disposition, thereby testing whether Hypothesis 1 holds true. Stepwise multiple regression automatically selects significant predictors, making it well suited for high-dimensional scenarios; by dynamically adjusting the set of variables, it improves the model accuracy. It effectively mitigates multicollinearity among predictors, reduces model errors, and enhances the predictive precision. By eliminating non-significant variables, stepwise regression yields a more parsimonious equation, boosting the computational efficiency and lowering the risk of overfitting (Ghani & Ahmad, 2010).
In line with Hypotheses 2 and 3, there may be mediating effects among self-regulated learning abilities, the use of AIGC for reflection, and critical thinking disposition, as well as among intrinsic learning motivation, the use of AIGC for reflection, and critical thinking disposition. To this end, this study conducts two mediating effect analyses using the Process plugin developed by Hayes (2012) in SPSS 26.0 (Statistical Package for the Social Sciences). Self-regulated learning ability and intrinsic learning motivation serve as the independent variables in the two mediating effect analyses, respectively, while the use of AIGC for reflection and critical thinking disposition act as the mediating and dependent variables in the analyses. If mediating effects in both analyses are established, then Hypotheses 2 and 3 are confirmed.

4. Results

4.1. Descriptive Statistics and Correlation Analysis of Questionnaire Data

The results regarding the descriptive statistics and correlation analysis of the sample are shown in Table 1. The data show that critical thinking disposition, AI literacy, the usage frequency of AIGC, the use of AIGC for reflection, self-regulated learning, and intrinsic learning motivation among the surveyed college students are all above the medium level (five-point scale, compared with 3 as the median). Gender and family economic status are below the medium level, indicating that the majority are female students and most students come from families with an average or relatively poor economic background. The dependent variable (critical thinking disposition) is significantly correlated with independent variables such as gender, family economic status, AI dependency, AI literacy, AIGC used for reflection, self-regulated learning, and intrinsic learning motivation (p < 0.05), but not with the usage frequency of AIGC (p > 0.05).

4.2. Multiple Stepwise Regression of Questionnaire Data

To test Hypothesis 1, we employed the multiple stepwise regression method to analyze the data, with the results shown in Table 2. Model 1 included gender, grade, family economic status, AI dependency, AI literacy, and the usage frequency of AIGC as independent variables for analysis. The results showed that only the coefficient for AI literacy was significant (T = 8.475, p < 0.001). Models 2, 3, and 4 successively incorporated the variables of AIGC used for reflection, self-regulated learning, and learning motivation into the regression model. The results indicated that the four variables—AI literacy, the use of AIGC for reflection, self-regulated learning, and learning motivation—were all significant across all models, suggesting that these variables have a significant impact on critical thinking disposition. Gender, family economic status, and AI dependency were only significant in some models, while grade and the usage frequency of AIGC were not significant in any of the models. This implies that these variables either have no impact or an insignificant impact on college students’ critical thinking disposition. The group of students who applied AIGC to reflective thinking exhibited a stronger critical thinking disposition, thereby validating Hypothesis 1.

4.3. Analysis of Mediation Effects of Questionnaire Data

According to Hypotheses 2 and 3, self-regulated learning and intrinsic learning motivation may serve as mediating variables between the use of AIGC for reflection and critical thinking disposition. To this end, this study used the Process plugin in SPSS 26.0 to conduct two mediation effect analyses, with the use of AIGC for reflection as the mediating variable, critical thinking disposition as the dependent variable, and self-regulated learning and intrinsic learning motivation as the independent variables. The results for Model 5 and Model 6 are shown in Table 3.
The results indicate that the use of AIGC for reflection mediates the relationship between self-regulated learning and critical thinking disposition, as well as between intrinsic learning motivation and critical thinking disposition (the confidence intervals do not include 0). In other words, students with stronger self-regulated learning abilities and intrinsic learning motivation are more inclined to use AIGC for reflection, which in turn enables them to better achieve the development of critical thinking skills. The mediating effects of self-regulated learning → AIGC use for reflection → critical thinking disposition and intrinsic learning motivation → AIGC use for reflection → critical thinking disposition are illustrated in Figure 1.

4.4. Analysis of the Interview Materials

The analysis of the interview materials reveals the following findings.
Firstly, the use of AIGC for reflection is more likely to enhance critical thinking skills. Some students reported that they typically experience improvements in critical thinking when using AIGC in the following scenarios: when dealing with complex problems and needing AIGC to provide different perspectives, they compare the strengths and weaknesses of these perspectives and analyze the underlying logic; when they find contradictions or irrationalities in the answers provided by AIGC, they conduct in-depth analyses of the answers; and when engaging in reflective thinking, they use AIGC to provide prompts and then think deeply based on these prompts. As one interviewed student noted, “When I first started using AIGC, I almost blindly trusted the answers it gave. After using it several times, I realized its responses could be flawed and shouldn’t be taken at face value. This forced me to change my approach: I now scrutinize and evaluate every answer it provides to see if there are any issues. Doing so has trained my critical thinking skills and taught me never to idolize anything—only what I carefully analyze and think through myself is truly reliable.” These processes prompt college students to think deeply about problems and examine and analyze issues and their underlying logic from different angles, thereby helping them to enhance their critical thinking skills.
Secondly, self-regulated learning abilities influence how college students use AIGC. Some students mentioned that they seek help from AIGC when formulating study plans, communicating with classmates, lacking motivation and passion for learning, or encountering learning difficulties. For instance, when faced with a challenging course, they use AIGC to filter learning materials and organize a framework of knowledge points, and they then analyze and reflect on the materials provided by the AIGC. Another example is as follows: when encountering a difficult problem, they use AIGC to break down the knowledge points and decompose the problem into several smaller problems. After solving the smaller problems, the larger problem becomes easily solvable. For example, one anonymous student remarked, “When I hit a learning roadblock, AIGC becomes my ‘on-demand tutor’—I ask it for problem-solving strategies and request recommendations for relevant learning resources.” This indicates that students with strong self-regulated learning abilities are adept in using AIGC to help with in-depth thinking to solve problems, rather than being at a loss, giving up, or directly copying the answers provided by AIGC.
Thirdly, intrinsic learning motivation affects how college students use AIGC. Some students stated that, in order to improve their own skills, they use AIGC to obtain deeper and broader knowledge. For example, they use AIGC to analyze cutting-edge academic trends and expand their professional horizons. They also use AIGC for topics that interest them. When learning Python programming, they ask AIGC to provide different answers to the same problem and gain a deeper understanding of programming knowledge through comparing different answers. For instance, one anonymous interviewee remarked, “When my Python code threw an error, I pasted the code into the AI. It not only pinpointed the bug but also suggested several optimizations.” This shows that students with strong intrinsic learning motivation are more inclined to use AIGC to improve their own skills and abilities and to internalize knowledge through independent exploration, rather than simply copying the answers provided by AIGC.

5. Discussion

5.1. Overview of Findings

Firstly, the usage frequency of AIGC is unrelated to critical thinking disposition, while the use of AIGC for reflection contributes to the development of critical thinking skills. The results of the correlation analysis show that, although the correlation between the usage frequency of AIGC and critical thinking disposition is negative and not significant (r = −0.04, p > 0.05), there is a significant correlation between the use of AIGC for reflection and critical thinking disposition (r = 0.32, p < 0.05). In the three regression analyses in Model 2, Model 3, and Model 4, the regression coefficients for the use of AIGC for reflection are all significant (p < 0.001). The interview results also show that using AIGC to analyze complex problems and compare different viewpoints and for reflective thinking can promote improvements in critical thinking skills. This indicates that using AIGC alone does not promote the development of critical thinking skills, but using AIGC in reflective thinking does. Therefore, it is concluded that Hypothesis 1 is valid.
Secondly, students with strong self-regulated learning abilities are more likely to use AIGC tools for reflective thinking and achieve better development in their critical thinking skills. The results of the mediation effect analysis show that self-regulated learning abilities have an indirect effect on critical thinking disposition through the use of AIGC for reflection (mediation effect value = 0.0510, the confidence interval does not include 0). The interview results show that students with stronger self-regulated learning abilities are more likely to use AIGC for problem analysis, problem solving, and reflective thinking activities. Combining the analysis of the questionnaire data and interview materials, it can be seen that students with strong self-regulated learning abilities are more inclined to use AIGC for reflective thinking. Since Hypothesis 1 is valid, it can be inferred that these students will benefit more from using AIGC than students with weak self-regulated learning abilities and achieve better development in their critical thinking skills. Therefore, it is concluded that Hypothesis 2 is valid.
Thirdly, students with strong intrinsic learning motivation are more likely to use AIGC tools for reflective thinking and achieve better development in their critical thinking skills. The results of the correlation and regression analyses show that the correlations between intrinsic learning motivation and critical thinking disposition, as well as between intrinsic learning motivation and the use of AIGC for reflection, are both significant (p < 0.001). The results of the mediation effect analysis show that intrinsic learning motivation has an indirect effect on critical thinking disposition through the use of AIGC for reflection (mediation effect value = 0.0266, the confidence interval does not include 0). The interview results show that students with stronger intrinsic learning motivation are more inclined to use AIGC to improve their abilities and knowledge reserves. Combining the analysis of the questionnaire data and interview materials, it is found that students with strong intrinsic learning motivation are more inclined to engage in reflective thinking when using AIGC. Compared with students with weak intrinsic learning motivation, they are more likely to achieve the development of their critical thinking skills when using AIGC. Therefore, it is concluded that Hypothesis 3 is valid.

5.2. Comparison with Previous Studies

Firstly, this study finds that the usage frequency of AIGC is unrelated to critical thinking disposition, but using it for reflective thinking can promote the development of critical thinking skills. This conclusion is distinct from those of several previous studies, yet it is essentially consistent. For instance, as mentioned earlier, Li et al. (2025) found that an increase in the frequency of using generative AI tools can enhance higher-order thinking skills, including critical thinking skills, which is different from the findings of this study. However, in reality, their study discovered that increasing the frequency of basic applications of AIGC (referring to tasks with low skill requirements, such as “searching for information” and “translating text”) does not affect higher-order thinking skills. In contrast, increasing the frequency of deep creative applications (referring to tasks with high skill requirements, such as “seeking inspiration or ideas for problem solving” and “assisting with homework”) can significantly improve higher-order thinking skills. Moreover, the interaction quality mediates the effect of the frequency of use on higher-order thinking skills. Clearly, this is essentially consistent with the findings of this study, as both indicate that using AIGC for higher-order thinking may help to enhance this skill.
Additionally, as previously stated, studies on the effectiveness of using AIGC in classroom teaching have found that its use can effectively promote the development of critical thinking skills, provided that college students apply it to the analysis and contemplation of learning issues. For example, Avello et al. (2024) randomly divided 41 Master’s students into two groups. The experimental group used ChatGPT as a personalized learning support tool to create characters, structure plots, and write dialog in digital storytelling, while the control group completed the same learning tasks without using ChatGPT. The study found that using AIGC could promote critical thinking skills by helping students in the experimental group to free up cognitive resources for the analysis and solution of complex problems (Avello et al., 2024). Similarly, Kofahi and Husain (2025) divided 53 undergraduates into two groups. In the experimental group, AIGC was used to provide personalized and interactive teaching to assist students in individual self-directed learning, while the control group engaged in teaching and learning activities in the traditional manner. The results showed that students in the experimental group had significant improvements in their self-perceived critical thinking skills and understanding of complex subjects. In these studies, students’ use of AIGC tools is shown to be a targeted educational engagement behavior, ensuring that they apply AIGC to the analysis and contemplation of learning issues. This is also essentially consistent with the findings of this study, i.e., both this study and the previously mentioned studies agree that only when AIGC is used for thinking can critical thinking skills be enhanced.
From a purely data-driven perspective, another possibility is that students with stronger critical thinking skills are more inclined to use AIGC for reflective thinking activities—seeking better answers rather than simply accepting those provided by AIGC. Empirical evidence already supports this view. For instance, Gerlich (2025) found, through survey data, that the higher the frequency of AI tool use, the lower individuals’ critical thinking skills. However, respondents with stronger critical thinking skills—who typically had higher educational attainment—were able to partially buffer the negative effects of AI use. They were more likely to cross-check AI-generated results and to leverage AI for reflective activities such as questioning, verifying, and refining, whereas individuals with lower educational levels tended to adopt the outputs directly, exhibiting greater “answer dependence”.
Secondly, this study finds that students with strong self-regulated learning abilities are more inclined to apply AIGC to reflective thinking, which is not entirely consistent with the conclusions of previous relevant studies, but it is essentially the same. Zhou et al. (2024) found that self-regulated learning abilities play a mediating role between the perceived ease of use of AIGC and critical thinking skills. When students find AIGC tools easy to use, they can more effectively employ self-regulation strategies, thereby promoting deeper cognitive processing. This study and the research by Zhou et al. (2024) reveal the role of self-regulated learning abilities in the use of AIGC from different perspectives, yet the results of both studies are essentially consistent—that is, self-regulated learning abilities affect the relationship between the use of AIGC and critical thinking skills, and this impact only occurs when students use AIGC tools for deep cognitive processing (reflective thinking).
Thirdly, this study finds that students with stronger intrinsic learning motivation are more inclined to apply AIGC to reflective thinking. There is currently no relevant research to directly corroborate this conclusion, but it is essentially consistent with the conclusions of existing relevant studies. For example, Lu et al. (2021) found that, in a collaborative inquiry-based learning environment, deep learning methods play an important mediating role between intrinsic motivation and higher-order thinking skills, and intrinsic learning motivation has a significant and positive impact on deep learning. Deep learning methods involve the in-depth processing of learning information and can predict understanding and critical thinking skills (Phan, 2006). Reflective thinking is essentially also the in-depth processing of learning information. Therefore, it can be stated that the conclusions of this study are essentially consistent with existing research.

5.3. Theoretical Explanations

Firstly, we consider the perspective of self-regulated learning theory. Self-regulated learning theory emphasizes that the learning of learners is a process of active monitoring, regulation, and reflection. Through this process, learners can plan, adjust, and monitor their cognitive, emotional, and behavioral responses to task demands (Panadero, 2017). This study found that students with strong self-regulated learning abilities are more adept in using AIGC tools to engage in reflective thinking, which can be explained from the perspective of self-regulated learning theory. Students with strong self-regulated learning abilities can clearly set learning goals and flexibly use external resources to optimize their learning strategies. AIGC, as an intelligent learning aid, can provide students with immediate feedback and conduct the multi-angle analysis and expansion of learning content, and students can use this to more accurately assess learning outcomes and adjust their learning strategies in a timely manner. This is the embodiment of the concepts of “metacognitive regulation” and “strategy execution” in self-regulated learning theory, proving that the synergy between AIGC and self-regulated learning abilities can effectively promote improvements in critical thinking skills.
Secondly, we consider the perspective of learning motivation theory. Learning motivation theory regards intrinsic learning motivation as the core driving force that propels learners to actively explore and engage in in-depth learning (Lu et al., 2021). This study found that students with strong intrinsic learning motivation are more inclined to apply AIGC to reflective thinking, which can be explained from the perspective of learning motivation theory. Intrinsic motivation drives students to pursue a deep understanding of knowledge and the enhancement of their own abilities. They regard learning as a process of self-actualization, rather than a task driven by external rewards. The open knowledge interaction and in-depth thinking guidance functions provided by AIGC perfectly meet the needs of these students for active learning. When students use AIGC to analyze problems from multiple dimensions, this process not only satisfies their intrinsic desire for knowledge exploration but also further strengthens their learning motivation through immediate cognitive challenges and feedback. This positive cycle encourages students to use AIGC more frequently to engage in reflective thinking, continuously break through thinking stereotypes, and thus achieve improvements in their critical thinking skills. This confirms the promoting effect of the combination of intrinsic learning motivation and AIGC tools on the cultivation of higher-order thinking skills.

5.4. Limitations and Future Research

This study has limitations in the following aspects. First, data were collected through self-reported questionnaires, which may involve response bias, but this study did not explore this bias. The sample was obtained from a university, and the size was limited. Second, the research was limited to open learning environments, and the findings are not fully applicable to structured classroom settings. Third, the exploration of mediating mechanisms focused on self-regulated learning and intrinsic motivation, failing to include other potentially influential variables (such as disciplinary differences and self-efficacy). Fourth, the results of the data analysis in this study reveal correlations among the investigated variables but do not imply any causal relationships; this caveat should be heeded by readers. Future research can be expanded through questionnaire surveys from more universities with a larger sample size and longitudinal studies to explore the dynamic impact of AIGC on critical thinking over time. Research methods capable of uncovering causal relationships among variables should also be employed in future studies to enhance the depth of the research.

5.5. Implications

Colleges and universities should actively construct a learning environment that supports the application of AIGC tools and encourage college students to use AIGC in reflective thinking. Relevant elective courses can be offered to help students to master the correct methods of using AIGC tools in professional learning and academic research. Meanwhile, learning resource platforms should be improved by integrating AIGC tools with case libraries for critical thinking training to provide students with more practical scenarios. University teachers need to transform their teaching concepts and guide students to apply AIGC tools in reflective thinking activities during classroom instruction. By designing collaborative and inquiry-based learning tasks, such as debates around contradictory viewpoints generated by AIGC, students can be prompted to take the initiative in reflection. In the teaching process, an emphasis should be placed on cultivating students’ self-regulated learning abilities, stimulating intrinsic learning motivation, assisting students in formulating study plans, monitoring the learning process, and arousing their interest in and curiosity about knowledge. College students should fully recognize the value of AIGC in cultivating critical thinking skills and actively integrate it into the learning process. AIGC tools should be consciously used to obtain diverse perspectives, verify ideas, and proactively seek feedback and suggestions for them. At the same time, they should strengthen the exercise of their self-regulated learning abilities, learn to plan their studies reasonably, actively mobilize intrinsic learning motivation, deepen their understanding of problems through interactions with AIGC, and achieve improvements in their critical thinking skills.

6. Conclusions

This study draws the following conclusions to answer the questions raised in the Introduction.
Firstly, in an open learning environment, the usage frequency of AIGC does not affect critical thinking skills, but applying AIGC to reflective thinking activities helps to enhance critical thinking. The essence of critical thinking lies in reflecting on thinking itself. When college students encounter phenomena or viewpoints that are contradictory to their inherent cognition, critical thinking is activated during the reflective thinking process on these viewpoints or phenomena. “Reflection” is the key to improving critical thinking skills. The role of AIGC is to provide viewpoints or factual materials that differ from students’ inherent cognition, creating conditions for the activation of critical thinking.
Secondly, the mechanism through which AIGC applied to reflective thinking activities promotes critical thinking skills is as follows: self-regulated learning and intrinsic learning motivation directly influence critical thinking and exert a mediating effect on critical thinking skills through the use of AIGC for reflection. Students with stronger self-regulated learning abilities or intrinsic learning motivation are more likely to use AIGC for reflective thinking activities and thus develop critical thinking skills.
Thirdly, from the perspectives of self-regulated learning theory and learning motivation theory, students with strong self-regulated learning abilities are adept in planning, monitoring, and adjusting the learning process. Students with strong intrinsic learning motivation, driven by interest and curiosity, are more likely to engage in proactive learning and use AIGC tools for such reflective thinking activities as sorting out ideas with AIGC’s help, verifying viewpoints, asking AIGC to identify logical flaws, obtaining diverse perspectives, and deepening problem analysis and understanding through interactions with AIGC. This process effectively enhances critical thinking skills and enables cognitive advancement.

Author Contributions

Conceptualization, W.Z. and X.L.; Methodology, W.Z.; Formal Analysis, W.Z.; Investigation, W.Z.; Data Curation, W.Z.; Writing—Original Draft Preparation, W.Z.; Writing—Review and Editing, W.Z.; Visualization, W.Z.; Supervision, W.Z.; Project Administration, W.Z.; Funding Acquisition, W.Z. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Hunan Province Education Science “14th Five Year Plan” 2023 Education Informatization Research Base (Technology Application Direction), Provincial Key Project in Education, grant number XJK23AJD052. The views expressed in this paper are the authors’ and do not necessarily represent the views of their institutions.

Institutional Review Board Statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the Academic Ethics Committee of the School of Education, Hunan University of Science and Technology (Ref. no. 2025 (02); 21 February 2025).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
CTDCritical thinking disposition
GraGrade
FESFamily economic status
AIDAI dependency
AILAI literacy
UFUsage frequency of AIGC
AURAIGC used for reflection
SRLSelf-regulated learning
ILMIntrinsic learning motivation

Appendix A

Participants’ self-regulated learning abilities, intrinsic learning motivation, critical thinking disposition, AI dependency, and AI literacy are assessed based on specific criteria. Participants rate each item on a 5-point Likert-type scale, which ranges from 1 (strongly disagree) to 5 (strongly agree). Below is a breakdown of these criteria.
  • Self-Regulated Learning Ability Scale
Goal setting
1. I have set standards for learning in each course.
2. I have set short-term (daily or weekly) goals as well as long-term goals (monthly or semester-long).
3. I have maintained high standards of learning in all subjects.
4. I set goals to help me manage my study time.
5. I do not let other matters lower the quality of my learning.
Environment structuring
6. I choose places to study where I won’t be easily distracted.
7. I find a comfortable place to study.
8. I know where I can study most effectively.
9. I choose a time to study when there are almost no distractions.
Task strategies
10. I take notes while studying because notes are important for learning.
11. When I get distracted during study, I use strategies to shift my attention back to studying.
12. I usually preview the course content before class.
13. In addition to what is taught in class, I assign myself some extra tasks to better grasp the course content.
Time management
14. I allocate extra time for my studies because I am aware that time is limited.
15. I try to schedule the same time for studying every day or week, and I stick to the timetable.
16. Although we don’t have to attend classes every day, I still try to distribute my study time evenly across each day.
Help seeking
17. When I encounter problems in my studies, I usually seek help from classmates or teachers.
18. I share the problems I encounter in my studies with my classmates.
19. For issues that arise during my studies, I try to communicate face-to-face with my classmates.
20. When faced with particularly thorny problems, I seek help from my teachers.
Self evaluation
21. I regularly summarize my learning situation and analyze any existing problems.
22. While studying, I reflect on whether my learning is effective.
23. I communicate with my classmates about our performance in learning.
24. I discuss with my classmates to identify differences between what I am learning and what they are learning.
  • Intrinsic Learning Motivation Scale
1. In such classes, I prefer course materials that truly challenge me so that I can learn something new.
2. In such classes, I prefer course materials that can arouse my curiosity, even if they are difficult to learn.
3. When I have the opportunity, I will choose course assignments that I can learn from, even if they do not guarantee good grades.
  • Critical Thinking Disposition Scale
1. During the learning process, I think about whether what I am learning is correct.
2. During the learning process, I judge the value of the new information or evidence that is presented to me.
3. I try to understand the content from different perspectives.
4. During the learning process, I evaluate different opinions to see which one is more reasonable.
5. As far as I know, I can identify facts that are supported by evidence.
  • AI Dependency Scale
1. Do I usually spend a lot of time thinking about how to use AI tools?
2. Do I increasingly want to use AI tools?
3. Do I use AI tools to forget about some of my personal troubles or dissatisfaction?
4. Have I tried to reduce my use of AI tools but failed?
5. Would I become restless or distressed if I were forbidden to use AI tools?
6. Has the overuse of AI tools had a negative impact on my work/study?
  • AI Literacy Scale
Awareness
1. I can distinguish between intelligent devices and non-intelligent devices.
2. I know that artificial intelligence technology can help me.
3. I can identify the artificial intelligence technology used in the applications and products I use.
Usages
4. I can skillfully use artificial intelligence applications or products to help me complete my daily work.
5. For me, learning to use new artificial intelligence applications or products is usually easy.
6. I can use artificial intelligence applications or products to improve my work efficiency.
Evaluation
7. After using for a while, I can evaluate the functions and limitations of artificial intelligence applications or products.
8. I can select the appropriate solution from the various solutions provided by intelligent agents.
9. I can choose the most suitable one from various AI applications or products for a specific task.
Ethics
10. When using artificial intelligence applications or products, I always adhere to ethical principles.
11. When using artificial intelligence applications or products, I remain vigilant about privacy and information security issues.
12. I always remain vigilant about the misuse of artificial intelligence technology.

References

  1. Andreassen, C. S., Torsheim, T., Brunborg, G. S., & Pallesen, S. (2012). Development of a facebook addiction scale. Psychological Reports, 110(2), 501–517. [Google Scholar] [CrossRef]
  2. Avello, R., Gajderowicz, T., & Gomez-Rodríguez, V. G. (2024). Is ChatGPT helpful for graduate students in acquiring knowledge about digital storytelling and reducing their cognitive load? An experiment. Revista de Educacion a Distancia (RED), 24(78), 604621. [Google Scholar] [CrossRef]
  3. Barnard, L., Lan, W. Y., To, Y. M., Paton, V. O., & Lai, S. L. (2009). Measuring self-regulation in online and blended learning environments. The Internet and Higher Education, 12(1), 1–6. [Google Scholar] [CrossRef]
  4. Bayuningsih, A. S., Usodo, B., & Subanti, S. (2017, September 18–19). Critical thinking level in geometry based on self-regulated learning. International Conference on Mathematics, Science and Education 2017 (ICMSE2017), Semarang, Indonesia. [Google Scholar]
  5. Cassidy, S. (2011). Self-regulated learning in higher education: Identifying key component processes. Studies in Higher Education, 36(8), 989–1000. [Google Scholar] [CrossRef]
  6. Chi, Q., Jiang, X., & Lv, Y. (2021, September 24–26). Analysis on the group differences of information literacy among local college students. 2021 4th International Conference on Information Systems and Computer Aided Education, Dalian, China. [Google Scholar]
  7. Essel, H. B., Vlachopoulos, D., Essuman, A. B., & Amankwa, J. O. (2024). ChatGPT effects on cognitive skills of undergraduate students: Receiving instant responses from AI-based conversational large language models (LLMs). Computers and Education: Artificial Intelligence, 6, 100198. [Google Scholar] [CrossRef]
  8. Facione, P. A., Sanchez, C. A., Facione, N. C., & Gainen, J. (1995). The disposition toward critical thinking. The Journal of General Education, 44(1), 1–25. [Google Scholar]
  9. Fadillah, M. A., Usmeldi, U., & Asrizal, A. (2024). The role of ChatGPT and higher-order thinking skills as predictors of physics inquiry. Journal of Baltic Science Education, 23(6), 1178–1192. [Google Scholar] [CrossRef]
  10. Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6. [Google Scholar] [CrossRef]
  11. Ghani, I. M. M., & Ahmad, S. (2010). Stepwise multiple regression method to forecast fish landing. Procedia-Social and Behavioral Sciences, 8, 549–554. [Google Scholar] [CrossRef]
  12. Gu, P., Wu, J., Cheng, Z., Xia, Y., Cheng, M., & Dong, Y. (2025). Scaffolding self-regulation in project-based programming learning through online collaborative diaries to promote computational thinking. Education and Information Technologies, 1–25. [Google Scholar] [CrossRef]
  13. Halpern, D. F. (1998). Teaching critical thinking for transfer across domains. Dispositions, skills, structure training, and metacognitive monitoring. American Psychologist, 53(4), 449–455. [Google Scholar] [CrossRef]
  14. Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling [White paper]. Available online: http://www.afhayes.com/public/process2012.pdf (accessed on 12 April 2025).
  15. Hooshyar, D., & Druzdzel, M. J. (2024). Memory-based dynamic bayesian networks for learner modeling: Towards early prediction of learners’ performance in computational thinking. Education Sciences, 14(8), 917. [Google Scholar] [CrossRef]
  16. Kofahi, M. A., & Husain, A. J. (2025). ChatGPT for operating systems: Higher-order thinking in focus. Journal of Information Technology Education: Research, 24, 1. [Google Scholar]
  17. Ku, K. Y., & Ho, I. T. (2010). Dispositional factors predicting Chinese students’ critical thinking performance. Personality and Individual Differences, 48(1), 54–58. [Google Scholar] [CrossRef]
  18. Li, M., Qiao, W., & Li, R. (2025). 大语言模型工具能促进高校学生的高阶思维能力发展吗? [Can big language modeling tools prom0ote the development of higher-order thinking abilities among college students?]. Journal of Modern Educational Technology 现代教育技术, 35(1), 34–43. (In Chinese). [Google Scholar]
  19. Lu, K., Pang, F., & Shadiev, R. (2021). Understanding the mediating effect of learning approach between learning factors and higher order thinking skills in collaborative inquiry-based learning. Educational Technology Research and Development, 69(5), 2475–2492. [Google Scholar] [CrossRef]
  20. Lu, K., Zhu, J., Pang, F., & Shadiev, R. (2025). Understanding the relationship between colleges students’ artificial intelligence literacy and higher order thinking skills using the 3p model: The mediating roles of behavioral engagement and peer interaction. Educational Technology Research and Development, 73(2), 693–716. [Google Scholar] [CrossRef]
  21. McPeck, J. E. (1981). Critical thinking and education. St. Martin’s Press. [Google Scholar]
  22. Menezes Cunha, N., Martinez, A., Kyllonen, P., & Gates, S. (2021). Cross-country comparability of a social-emotional skills assessment designed for youth in low-resource environments. International Journal of Testing, 21(3–4), 182–219. [Google Scholar] [CrossRef]
  23. Mogavi, R. H., Deng, C., Kim, J. J., Zhou, P., Kwon, Y. D., Metwally, A. H. S., Tlili, A., Bassanelli, S., Bucchiarone, A., Gujar, S., Nacke, L. E., & Hui, P. (2024). ChatGPT in education: A blessing or a curse? A qualitative study exploring early adopters’ utilization and perceptions. Computers in Human Behavior: Artificial Humans, 2(1), 100027. [Google Scholar] [CrossRef]
  24. Okonkwo, C. W., & Ade-Ibijola, A. (2021). Chatbots applications in education: A systematic review. Computers and Education: Artificial Intelligence, 2, 100033. [Google Scholar] [CrossRef]
  25. Panadero, E. (2017). A review of self-regulated learning: Six models and four directions for research. Frontiers in Psychology, 8, 422. [Google Scholar] [CrossRef]
  26. Phan, H. P. (2006). Examination of student learning approaches, reflective thinking, and epistemological beliefs: A latent variables approach. Electronic Journal of Research in Educational Psychology, 4(3), 577–610. [Google Scholar]
  27. Shi, B., & Shen, J. (2007). 家庭社会经济地位、智力和内部动机与创造性的关系. [The relationship between family socioeconomic status, intelligence, internal motivation, and creativity]. Psychological Development and Education 心理发展与教育, 1(1), 30–34. (In Chinese). [Google Scholar]
  28. Wang, L., & Chen, M. (2010). The effects of game strategy and preference-matching on flow experience and programming performance in game-based learning. Innovations in Education and Teaching International, 47(1), 39–52. [Google Scholar] [CrossRef]
  29. Widana, I. W. (2022). Meta-analysis: The relationship between self-regulated learning and mathematical critical reasoning. Education Innovation Diversity, 1(4), 64–75. [Google Scholar] [CrossRef]
  30. Xu, Y., Zhu, J., Wang, M., Qian, F., Yang, Y., & Zhang, J. (2024). The impact of a digital game-based AI chatbot on students’ academic performance, higher-order thinking, and behavioral patterns in an information technology curriculum. Applied Sciences, 14(15), 6418. [Google Scholar] [CrossRef]
  31. Yu, H. (2023). Reflection on whether Chat GPT should be banned by academia from the perspective of education and teaching. Frontiers in Psychology, 14, 1181712. [Google Scholar] [CrossRef]
  32. Zhang, S., Zhao, X., Zhou, T., & Kim, J. H. (2024). Do you have ai dependency? the roles of academic self-efficacy, academic stress, and performance expectations on problematic ai usage behavior. International Journal of Educational Technology in Higher Education, 21(1), 34. [Google Scholar] [CrossRef]
  33. Zhou, X., Teng, D., & Al-Samarraie, H. (2024). The mediating role of generative ai self-regulation on students’ critical thinking and problem-solving. Education Sciences, 14(12), 1302. [Google Scholar] [CrossRef]
  34. Zhu, J., Xu, L., & Ma, J. (2025). 生成式人工智能如何赋能学生学习—基于大学生自我调节学习行为的实证研究. [How generative artificial intelligence empowers student learning: An empirical study based on self regulated learning behavior of college students]. Research on Higher Engineering Education 高等工程教育研究, 1(2), 66–72. (In Chinese). [Google Scholar]
Figure 1. Intermediary effect diagram.
Figure 1. Intermediary effect diagram.
Education 15 00977 g001
Table 1. Sample descriptive statistical analysis and correlation analysis results.
Table 1. Sample descriptive statistical analysis and correlation analysis results.
VariableMS.E.123456789
CTD3.790.43---------
Sex0.250.430.09 *-- -----
Gra2.600.800.44−0.23 ***-------
FES2.750.620.09 *0.04−0.05------
AID2.980.740.09 *0.060.06−0.01-----
AIL3.870.490.35 ***0.01−0.010.060.25 ***----
UF3.281.99−0.04−0.07−0.03 ***−0.060.07−0.02---
AUR3.720.890.32 ***0.02−0.02−0.050.15 ***0.30 ***−0.02--
SRL3.420.580.48 ***−0.060.080.14 *0.30 ***0.05 ***−0.060.25 ***-
ILM3.400.850.43 ***0.16 ***−0.10 *0.050.15 ***0.27 ***−0.060.16 ***0.44 ***
Notes: * p < 0.05, *** p < 0.001. CTD, Gra, FES, AID, AIL, UF, AUR, SRL, and ILM are abbreviations for critical thinking disposition, grade, family economic status, AI dependency, AI literacy, usage frequency of AIGC, AIGC use for reflection, self-regulated learning, and intrinsic learning motivation, respectively. The same applies below.
Table 2. Regression analysis of the impact of AIGC usage on critical thinking tendencies of college students.
Table 2. Regression analysis of the impact of AIGC usage on critical thinking tendencies of college students.
Model 1Model 2Model 3Model 4
Sex0.096 (1.830)0.094 (1.850)0.134 *** (2.843)0.088 (1.897)
Gra−0.011 (−0.370)−0.006 (−0.203)−0.019 (−0.742)−0.002 (−0.072)
FES0.056 (1.574)0.070 * (2.026)0.024 (0.741)0.028 (0.913)
AID0.001 (0.029)−0.013 (−0.444)−0.077 *** (−2.703)−0.077 *** (−2.775)
AIL0.393 *** (8.475)0.315 *** (6.717)0.173 *** (3.791)0.149 *** (3.357)
UF−0.009 (−0.807)−0.007 (−0.626)−0.002 (−0.162)0.001 (0.136)
AUR-0.149 *** (5.948)0.116 *** (4.96)0.113 *** (4.974)
SRL--0.385 *** (9.984)0.291 *** (7.127)
ILM---0.153 *** (5.993)
Constant2.210 *** (9.842)1.941 *** (8.720)1.621 *** (7.791)1.467 *** (7.212)
N566566566566
Notes: * p < 0.05, *** p < 0.001. The data in the table are the non-standardized coefficients of the variables, and the data in parentheses are the T-values; the same applies below.
Table 3. Analysis of the mediating effects of self-regulated learning, intrinsic learning motivation, and the use of AIGC for reflection on the critical thinking tendencies of college students.
Table 3. Analysis of the mediating effects of self-regulated learning, intrinsic learning motivation, and the use of AIGC for reflection on the critical thinking tendencies of college students.
Model 5PathEffect ValueConfidence Interval (95%)Proportion of Effect
Direct effectSRL → CTD0.4036[0.334, 0.473]88.78%
Mediation EffectSRL → AUR → CTD0.0510[0.0257, 0.0801]11.22%
Total effect 0.4546
Model 6PathEffect ValueConfidence Interval (95%)Proportion of Effect
Direct effectILM → CTD0.2529[0.2057, 0.3000]90.48%
Mediation EffectILM → AUR → CTD0.0266[0.0110, 0.0454]9.52%
Total effect 0.2795
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhang, W.; Liu, X. Artificial Intelligence-Generated Content Empowers College Students’ Critical Thinking Skills: What, How, and Why. Educ. Sci. 2025, 15, 977. https://doi.org/10.3390/educsci15080977

AMA Style

Zhang W, Liu X. Artificial Intelligence-Generated Content Empowers College Students’ Critical Thinking Skills: What, How, and Why. Education Sciences. 2025; 15(8):977. https://doi.org/10.3390/educsci15080977

Chicago/Turabian Style

Zhang, Weiping, and Xinxin Liu. 2025. "Artificial Intelligence-Generated Content Empowers College Students’ Critical Thinking Skills: What, How, and Why" Education Sciences 15, no. 8: 977. https://doi.org/10.3390/educsci15080977

APA Style

Zhang, W., & Liu, X. (2025). Artificial Intelligence-Generated Content Empowers College Students’ Critical Thinking Skills: What, How, and Why. Education Sciences, 15(8), 977. https://doi.org/10.3390/educsci15080977

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop