By reshaping media organizational structures, operational practices, and core tasks, AI-based media production is becoming increasingly prevalent. This transformation in how users interact with and consume media will bring disruptive and challenging influences to the future media industry (Carlson, 2015). The technical complexity and diversity of media production in the AI era highlight the urgency of reforming traditional media education (Zhang, 2022a). Baptista et al. (2025) argue that the rapidly developing “intelligent +” media industry imposes new requirements on media professionals, who now need to possess interdisciplinary knowledge, expertise in intelligent technologies, data analysis skills, and the ability to create content quickly. Pavlik (2023) advocates for the establishment of interactive media education systems based on AI technologies to guide students in effectively collaborating with AI.

However, in current media education, the integration of artificial intelligence still tends to focus primarily on technological development, neglecting the cultivation of students' individual intelligent thinking and professional confidence (Li et al., 2022). As a result, educational methods lag behind the evolving needs of the intelligent media industry (Hu et al., 2018). Furthermore, the failure to foster students' intrinsic motivation and proactive engagement in the acceptance of intelligent technologies has led to suboptimal teaching outcomes (Zhang, 2022b). Liu (2024) argues that, in the context of media education, it is crucial to accurately assess students' individual characteristics, analyze their behaviors, and address their personal awareness and attitudes. Niu et al. (2024) suggest that exploring the factors influencing students' willingness to adopt intelligent technologies will allow for the development of targeted strategies. Research by Tahat et al. (2023) analyzed the willingness of media students to use digital technology based on the Technology Acceptance Model (TAM), highlighting that perceived usefulness and perceived ease of use remain critical factors affecting ICT acceptance. Additionally, Soliman et al. (2025) emphasize the importance of considering students' personal attitudes toward the use of generative AI (GenAI) tools and creating interactive learning experiences to help students become well-informed citizens in an AI-driven world.

Teo et al. (2011) were the first to suggest that high levels of mindfulness promote organizational adoption of technology and enhance the efficiency of its application. Poston and Kettinger (2014) argue that mindfulness not only influences how new technologies are introduced, but also improves user satisfaction and the ability to understand technology. Sun et al. (2016) assert that a user's state of mindfulness is a critical factor in determining whether technology can be used consistently after initial exposure. Pflügner and Maier (2020) found that employees with higher levels of mindfulness experienced lower levels of technology-related stress, including overload, complexity, intrusiveness, and uncertainty. A recent study by Jang et al. (2021) found that mindfulness has a direct positive impact on both technology use and job performance in technology-intensive environments. For college students, mindfulness is a key factor in understanding their willingness to accept technology (Corti and Gelati, 2020). Mindfulness enhances students' creative learning and innovative acceptance, which, in turn, increases their willingness to use technology (Yeh et al., 2019).

With the development and widespread adoption of artificial intelligence, research on the specific impact of an individual's mindfulness level on the acceptance of intelligent technology has gradually emerged (Chuang et al., 2024). Hurst (2021) points out that providing users with personalized mindfulness experiences can effectively enhance digital skills in applications and tools; thus, digital mindfulness has been proposed as a mental health intervention and applied with increasing urgency. Rohwer et al. (2022) argue that technological stress is becoming more prevalent in the digitization of work, and both mindfulness and mindfulness in technology use positively affect the reduction of technological stress. Gupta et al. (2023) advocate for the development of mindfulness-based digital competency frameworks to support the sustainability of digital capabilities. Chen et al. (2024) confirmed the effect of mindfulness on emotional orientation and the intention to continue using Chat-GPT both during and after its application.

The Technology Acceptance Model (TAM), first proposed by Davis in 1989, is a theory of rational behavior, self-efficacy, and expectation confirmation, designed to explain how individuals adopt and use new technologies, particularly in the workplace. Davis et al. extended TAM by incorporating behavioral intentions as a component, which enhanced its explanatory power regarding user behavior. In subsequent research, Venkatesh further extended the TAM framework, developing the TAM2 model in 2000 and the Unified Theory of Technology Acceptance and Use (UTAUT) in 2003. By now, TAM remains one of the foundational theories for studying individuals' intentions and behaviors related to technology acceptance and usage, as it provides a detailed explanation of decision-making processes in technology adoption.

Recent studies have shown that the theoretical frameworks of research on college students' use or rejection of AI technology are primarily based on extensions of the classical TAM (Roy et al., 2022; Chen et al., 2024; Bekturova et al., 2025). Soliman et al. (2023) analyze the factors influencing the use of metaverse learning platforms by integrating TAM and Self-Determination Theory (SDT), suggesting that these factors significantly affect the adoption of metaverse platforms. Dahri et al. (2025) examine college students' willingness to use AI technology by combining SDT and TAM. Ma (2025) integrates personal innovativeness with TAM to explore the key factors affecting Chinese postgraduate students' willingness to use Chat-GPT, confirming that perceived usefulness (PU), perceived ease of use (PEU), and personal innovativeness (PI) have positive effects. Zhang et al. (2025) integrate TAM with Expectation-Confirmation Theory (ECT) and the Theory of Reasoned Action (TRA) into a comprehensive model, finding that performance expectation (PE) and technological innovation readiness index (TRII) have a significant positive impact on postgraduates' willingness to continue using AI. Al-Zahrani and Alasmari (2025) investigate middle school students' willingness to use AI in universities across the Middle East and North Africa region. Guided by TAM and the Diffusion of Innovations (DoI) theory, they find that PU, PEU, and personalized learning are significantly correlated with students' intention.

Based on research in media education, mindfulness, and the acceptance of AI, it has been confirmed that mindfulness plays a crucial role in the process of individuals' adoption of intelligent technology. However, most studies have focused on identifying the elements of mindfulness that affect technology acceptance, and there is a lack of a systematic theoretical framework and model construction. The investigation of mindfulness within the Technology Acceptance Model (TAM) is still in an exploratory phase. These researches are not only limited in number but also relatively simplistic in approach, often focusing on a one-dimensional, linear mechanism through which mindfulness influences the intention to use technology. Furthermore, studies examining how mindfulness impacts the college student's adoption of intelligent technology, particularly the students majoring in media, are still with a notable absence of systematic empirical studies.

Given the lack of a comprehensive framework across the three subjects including media education, mindfulness, and AI technology adoption, this study aims to explore whether and how the level of mindfulness affects media students' intention to accept and use intelligent technology, utilizing the mindfulness scale and the classical TAM. First, the role and function of mindfulness will be considered. Second, the internal structural mechanisms of mindfulness in relation to technology adoption will be examined. The main research questions are as follows:

RQ1: Does mindfulness affect media students' intention to adopt AI technology?

RQ2: If so, what are the factors and mechanisms by which mindfulness affects media students' intention? Are there other influencing factors?

In answering these questions, this paper focuses on how to enhance the internal drive of media students to use technological skills through the training and cultivation of mindfulness.

The Technology Acceptance Model (TAM) includes several key variables, such as perceived usefulness (PU), perceived ease of use (PEU), behavioral intention (BI), and other external factors (Venkatesh and Davis, 1996). Among these, PU and PEU are the core variables, effectively explaining users' perceptions of new technology. PU refers to the user's belief that using the technology can improve performance, while PEU refers to the user's perception of how easy the technology is to use (Kim et al., 2009; Bischof-dos-Santos et al., 2017). Given the robustness of TAM, subsequent studies have largely extended the model, combining various research subjects and contexts to explain users' acceptance of new technologies (Bouaguel and Alsulimani, 2022). Recently, scholars have added personal innovativeness (PI) and perceived risk (PR) to TAM to better explain students' acceptance of AI technologies (Yu et al., 2024b; Liu et al., 2024). Lee et al. (2021) confirmed that PI is positively associated with students' intention to use AI, while Pan et al. (2024) explored the role of PR, finding that it moderates the relationship between PU and PEU The integration of factors such as PI and PR not only extends the classical TAM model but also provides more targeted insights into the adoption of artificial intelligence (Almeida et al., 2025).

Mindfulness plays a key role in addressing negative emotions, cognitive challenges, and improving cognitive flexibility (Vogus and Sutcliffe, 2012; Yang et al., 2024). Moreover, mindfulness increasingly influences the analysis of psychological characteristics such as individuals' mindsets, values, beliefs, and emotions Wamsler and Brink, (2018). Recent studies have integrated mindfulness into the analysis of individuals' perception, application, and evaluation of new technologies (e.g., Oeldorf-Hirsch and Chen, 2022). Through this integration with the Technology Acceptance Model (TAM), it has been shown that mindfulness significantly impacts perceived usefulness, ease of use, and actual willingness to use technology. Specifically, it influences the cognitive processes of understanding, perceiving, and implementing AI technology, improving users' perceptions of its feasibility and accessibility. By strengthening the psychological influence, mindfulness promotes the usability, ease of use perception, and actual use behavior of AI technology (Tan, 2025).

This study systematically examines the impact of mindfulness levels on media students' intention to use artificial intelligence (AI) technology by integrating mindfulness into the classical Technology Acceptance Model (TAM). It explores the relationship between mindfulness levels and the intention to use AI technology, proposing that mindfulness significantly influences perceived usefulness (PU), perceived ease of use (PEU), and the actual intention to use technology. Additionally, by incorporating personal innovativeness (PI) and perceived risk (PR) into TAM, the study identifies other important factors that predict students' intention to adopt AI technology. Figure 1 illustrates the theoretical model constructed in this study.

The participants in this study were Chinese media students, including students from the majors in journalism and communication, television and film, and digital media art at both the undergraduate and graduate levels. Compared with the use of AI in majors like engineering and medicine, the adoption of AI technologies such as multimodal content generation, automated editing, and content creation has introduced a new generation of the media industry. These advancements bring about fundamental changes and even disruptive transformations (Nguyen and Hekman, 2022). Furthermore, when media students use artificial intelligence technologies, they must heavily rely on their perceptions and evaluations of the technology's usefulness, controllability, and value to ensure the ethical integrity and content accuracy of the generated material (Newman, 2022). As a result, media students face higher psychological complexity and situational sensitivity compared to students from other majors when using AI tools.

The minimum sample size for this study was determined using the G^*Power tool (Faul et al., 2009). With parameters set to a moderate effect size of 0.15, an error rate of 0.05, an effect strength of 0.8, and 7 predictors, the minimum sample size required for credible results was 103. The sample was selected using stratified random sampling, with strata based primarily on participants' major and academic year, ensuring a representative distribution of students across the study's relevant subgroups.

This study utilized the Questionnaire Star platform to design and distribute the questionnaire in March 2025. The distribution was conducted offline, with face-to-face administration, on-site completion, and collection. Of the 610 returned questionnaires, invalid responses were identified and excluded based on the following criteria: (1) key items were missing or unanswered, and (2) the response time was significantly shorter than the average time (i.e., < 1 min). After excluding 21 invalid responses, 588 valid questionnaires remained, yielding a response rate of 96.6%. For data analysis, SmartPLS (V4.1) was used to assess the direct and indirect effects among variables in the structural equation model, while the Process macro in SPSS 26 was applied for moderation analysis.

The questionnaire consists of two sections: the first section gathers basic demographic information from participants, and the second section assesses the variables in the model, including mindfulness, perceived usefulness, perceived ease of use, personal innovativeness, perceived risk, willingness to use, and actual use. A Likert scale was used to measure these variables. The specific test items are as follows:

The gender distribution of the survey sample is as follows (see Table 2): 22% male, with a total of 130 males, and 78% female, with a total of 458 females. Regarding the educational level, 84% of the participants are undergraduates, with a total of 493 students. In terms of majors, the distribution is as follows: 241 students majoring in Drama and Film, 159 students in Digital Media Arts, and 188 students in Journalism and Communication.

To assess model reliability, composite reliability (CR) and average variance extracted (AVE) were calculated. Table 3 shows CR values ranging from 0.838 to 0.935, all exceeding the recommended threshold of 0.70 (Urbach and Ahlemann, 2010). Most AVE values were above the recommended level of 0.50, with the AVE for mindfulness slightly below the threshold at 0.488. However, it still met the criteria when considered alongside its CR value of 0.934.

To further assess the convergent validity of the structural equation model, this study conducted a supplementary analysis using the square root of the AVE values. The results (see Table 4) show that the square root of the AVE values for all first-order variables are higher than the correlation coefficients between each variable and the other variables. This further indicates that discriminant validity is good.

An HTMT analysis of the scales was conducted using SmartPLS. The HTMT ratio values obtained (see Table 5) indicate that all the variables have HTMT values < 0.85, suggesting that the variables exhibit good discriminant validity statistically. This means that the variables are conceptually independent and can individually explain different aspects of the research. Therefore, from the perspective of discriminant validity, these variables are both valid and reliable for use in the study.

The data collected through the survey may be subject to common method bias. To address this issue, a multicollinearity test was performed. The six factors extracted for this study were analyzed for collinearity. The results (see Table 6) indicate that all variables had VIF values lower than 3.3. According to Kock's perspective, a VIF >3.3 is considered indicative of pathological multicollinearity, which could be affected by common method bias and contamination. This finding suggests that the measurements in this study are not significantly affected by covariance issues.

Table 7 presents the results of the theoretical effect test for the model in this study. R² is a measure of the explanatory power of the endogenous variables. An R² >0.67 indicates strong explanatory power; an R² between 0.33 and 0.67 indicates moderate explanatory power; and an R² < 0.19 indicates weak explanatory power. Although the R² values for the variables PI (Personal Innovativeness) and PEU (Perceived Ease of Use) are below 0.19, the model's SRMR (Standardized Root Mean Square Residual) is 0.058 (with SRMR < 0.8 indicating an acceptable model fit), and the NFI (Normed Fit Index) is 0.834 (with NFI closer to 1 indicating a better fit). Therefore, the structural equation model constructed in this study still possesses a certain level of explanatory power and persuasiveness.

Using SmartPLS and SPSS29 software, the constructed model framework of this study underwent tests for direct effects, indirect path effects, total effects, and moderating effects. The final structural equation model, along with path coefficients and significance indices for each path, are shown in Figure 2.

Table 8 presents the variance, confidence intervals, T-values, and P-values (significance) for each path, calculated using the SmartPLS bootstrap method, with 5,000 re-samplings. From the data analysis, the relationships between mindfulness, perceived usefulness, perceived ease of use, personal innovativeness, and actual usage intention of artificial intelligence (AI) among media major students are clearly identified.

Table 8 shows the variance, confidence interval, t-value, and p-value (significance) for each path, calculated using the SmartPLS bootstrap method and repeated 5,000 times. Therefore, through data analysis, the relationship between perceived usefulness, ease of use, personal innovation, and actual willingness to use artificial intelligence by college students majoring in mindful media can be clearly understood. Data analysis shows that the path coefficient of mindfulness on perceived usefulness is p = 0.002, p < 0.01 (SD = 0.043, t = 3.165), indicating that mindfulness has a significant, positive effect on perceived usefulness. In the impact of mindfulness on perceived ease of use, the path coefficient is p = 0.009, p < 0.01 (SD = 0.046, t = 2.597), indicating that mindfulness has a significant, positive effect on perceived ease of use. In the impact of mindfulness on the willingness to use artificial intelligence technology, the path coefficient is p = 0, p < 0.001 (SD = 0.036, t = 4.175), indicating that mindfulness has a significant, positive impact on the willingness to use artificial intelligence. In the impact of mindfulness on personal innovation, the path coefficient is p = 0.000, p < 0.001 (SD = 0.043, t = 4.949), indicating that mindfulness will have a significant, positive effect on personal innovation. In the impact of personal innovation on the actual use of artificial intelligence, the path coefficient is p = 0, p < 0.001 (SD = 0.046, t = 10.4), indicating that personal innovation has a significant, positive impact on the actual behavioral intention to use intelligent technology. In addition, the path coefficient p = 0, p < 0.001 (SD = 0.019, t = 3.785) in the process in which individual innovativeness exists as a mediator variable, and the path coefficient p = 0, p < 0.001 (SD = 0.019, t = 3.785), expresses that personal innovativeness does serve as an intermediate effect of mindfulness on the actual behavioral intention to use AI, and that mindfulness can moderate this effect by improving students' personal innovativeness, and successfully promote the acceptance and use of artificial intelligence tools. In the impact of perceived usefulness on the actual behavioral intention of artificial intelligence, the path coefficient p = 0, p < 0.001 (SD = 0.05, t = 6.9) indicates that perceived usefulness has a significant positive impact on the actual behavioral intention to use artificial intelligence. In the impact of perceived ease of use on the actual behavioral intention to use artificial intelligence, the path coefficient is p = 0.085, p > 0.05 (SD = 0.051, t = 1.721), indicating that perceived ease of use does not have a significant effect on the actual behavioral intention to use artificial intelligence.

Finally, the H1, H2, H3, H4, H5, H6, H7, H9, and H10 hypotheses proposed in this study are effectively supported, while H8 and H11 are not effectively verified.

The Process macro in SPSS 29 was used to test the moderating effect, with the results presented in Table 9. Additionally, simple adjustment slope charts were created using Excel software, as shown in Figures 3, 4. The results reveal that perceived risk has a significant negative moderating effect on the relationship between perceived usefulness and actual intention to use (t = −3.4502, p = 0.0006, < 0.001). Perceived risk also significantly moderates the relationship between personal innovativeness and actual use intention (t = −6.2721, p = 0.000, < 0.001). However, perceived risk does not significantly moderate the relationship between perceived ease of use and actual usage intention (t = −0.513, p = 0.9591).

The research indicates that mindfulness has a significant positive impact on the perceived usefulness (PU) of intelligent tools among media students. This supports Hypothesis 1, aligning with the findings of Liebherr et al. (2024) and Dai et al. (2022). Mindfulness, by promoting a focus on the present moment, enhances the ability to organize and analyze problems clearly. It also facilitates the secretion of dopamine and promotes wellbeing, which buffers against external distractions (Marsh et al., 2024), thereby increasing individuals' perceived usefulness of technology. Similarly, Hypothesis 2 is also supported, showing that the higher the level of mindfulness, the stronger the perceived ease of use of AI tools. This finding is consistent with previous studies (Ameer and Zubair, 2020; Liu et al., 2025), which suggest a positive spiral between mindfulness and emotions. As mindfulness levels increase, individuals experience more positive emotions, which help them overcome technology-related anxiety. This, in turn, leads to a more active evaluation of the technology's utility and ultimately enhances the perceived ease of use. Finally, Hypothesis 3 is also effectively supported. The findings indicate that mindfulness plays a crucial role in enhancing students' actual intention to use technology, a conclusion that resonates with studies by Galante et al. (2016) and Hämäläinen et al. (2021).

This study supports Hypothesis 4, providing empirical evidence for a significant, positive relationship between mindfulness and personal innovativeness. This aligns with the findings of Jobbehdar Nourafkan et al. (2022) suggest that individuals with higher levels of mindfulness tend to exhibit greater creativity, which, in turn, drives students to engage more consciously in cultivating personal creative abilities. Additionally, Henriksen et al. (2020) assert that mindfulness training, regardless of duration, enhances personal innovativeness by fostering creativity. Furthermore, Hypothesis 5 of this study demonstrates a significant positive relationship between personal innovativeness and students' intention to use AI technologies. This finding is supported by previous research (Kasilingam, 2020; Sowmya et al., 2023), which shows that personal innovativeness, as an innate trait, drives students to adopt new technologies more readily. This aligns with Strzelecki et al. (2024), who suggest that individuals exhibiting higher personal innovativeness exhibit behaviors that reduce the anxiety and stress typically associated with new technologies like AI, thus promoting acceptance and enhancing the intention to use AI tools (Kwarteng et al., 2023).

This study assumes effective support for H6. This suggests that mindfulness not only directly influences actual intention but also facilitates the adoption of new technologies by enhancing personal innovativeness. High levels of mindfulness improve cognitive abilities, such as memory and attention, which foster innovative thinking and maintain innovation-driven behaviors, leading to a greater willingness to adopt and use new technologies during learning tasks (Golubickis et al., 2024). Conversely, individuals with low levels of mindfulness find it more challenging to recognize and adjust to changes in their emotional states. When in a negative emotional state, their personal innovativeness is diminished, which ultimately decreases their willingness to use AI technologies (Petrou and Jongerling, 2022).

The findings of this study not only reinforce the central role of perceived usefulness in predicting actual use intention, as established in previous TAM research, but also underscore its significance in shaping users' technology adoption behavior. Specifically, students' perception of the usefulness of artificial intelligence tools enhances their expectation confirmation, leading to the belief that these tools are effective in task processing (Niu and Mvondo, 2024), which in turn increases their willingness to use them (Bilquise et al., 2023). When students perceive AI tools as highly useful, concerns related to AI ethical risks, privacy issues, and negative evaluations are mitigated (Camilleri, 2024), which strengthens their expectancy and self-efficacy. By reducing emotional barriers such as anxiety or worry, the perceived effectiveness of the tools is enhanced, ultimately fostering a greater willingness to use artificial intelligence technologies (Acosta-Enriquez et al., 2024).

Students' perceived ease of use of artificial intelligence technology does not significantly impact their actual usage intention, and research hypothesis 8 has not been effectively supported. Similarly, Soliman et al. (2024) found that perceived ease of use was not significantly related to the actual intention to use it. Zhao et al. (2025) also suggest that students' intentions may be influenced by various factors, including technology compatibility, facilitating conditions, and trust, which are more directly associated with behavior. They argue that focusing solely on perceived ease of use does not sufficiently explain actual usage intention. Recent studies by Oved and Alt (2025) also indicate that perceived ease of use is not the sole influencing factor; other external contexts, such as cultural, social, and technological environments, play a critical role in shaping users' actual intention to use.

This study hypothesizes that H9 is effectively supported, indicating that perceived risk negatively moderates the relationship between individual innovativeness and actual usage intention. Perceived risk is a crucial factor, primarily due to students' fear and anxiety regarding the perceived rationality and safety of AI technology, which creates uncertainty in their cognition and application of intelligent technology (Ye et al., 2024). This heightened perception of risk exacerbates students' negative emotions, such as panic, worry, and anxiety, before they engage with the technology. As a result, it increases the pressure associated with using AI technology and inadvertently weakens the relationship between personal innovativeness and the actual intention to use it (Bullock et al., 2025).

This study also supports hypothesis H10, which posits that perceived risk negatively moderates the relationship between perceived usefulness and actual intention to use. Perceived risk exacerbates the distrust of AI technology among Chinese college students majoring in Media (Khan and Khan, 2019; Chung and Kwon, 2025). This distrust stems from negative assumptions regarding the use of AI technologies, including potential fears about their accuracy and effectiveness in smart education contexts (Wu et al., 2023). Additionally, Duffett et al. (2024) found that the higher the perceived risk of technology, the lower the positive impact of individual innovativeness on the intention to use it.

However, hypothesis H11 is not supported in this study, indicating that perceived risk does not significantly moderate the relationship between perceived ease of use and actual usage intention. This result is consistent with the findings of Li et al. (2024), who similarly found no significant moderating effect of perceived risk in this relationship. One possible explanation is that perceived ease of use is largely influenced by users' emotional experiences, such as feelings of depression, joy, or frustration, which are shaped more by individual emotions than by perceived risk (Mohamad et al., 2021). Similar conclusions have been drawn by Lei et al. (2022), Tilibasa et al. (2023), and Goel et al. (2025), who assert that the relationship between perceived ease of use and actual usage intention is more complex than previously assumed. Moreover, the high level of technological familiarity within the sample may have introduced bias, potentially limiting the model's capacity to detect significant moderating effects.