2.1 Research status

AIGC is a breakthrough technological development. Tencent Research Institute (2023) [11] believes that AIGC (Artificial Intelligence Generated Content) models such as Stable Diffusion, DALL-E2 and MidJourney have sparked a craze in the field of AI mapping. The emergence of these models marks the deep penetration of AI into the field of art.

Through previous research [12], many scholars have begun to pay attention to the development of AIGC in the field of art. Current research is mainly divided into four areas:

(1) Verifying the advantages of AIGC technology. Most related studies use empirical analysis to measure the effects of AIGC technology on various aspects, such as design efficiency [13], creative expression [14], design interaction [15,16], creative outcomes [17], design strategy and innovation confidence [18]. In particular, there are many studies in specific fields such as fine arts and product design;

(2) Exploring the risks of AIGC technology. The explosive growth of AIGC in a short period of time has led to shortcomings in the standardised use of AIGC and has also concealed many hidden dangers. Elali & Rachid (2023) [19] stated that the use of AIGC tools involved in the design process can create risks of infringement and piracy. Dugan et al. (2023) [20] stated in their article that the use of AIGC tools in the design process provides false and incorrect information, and also stated in their research that artificial intelligence can promote violent information and spread incorrect content. Finally, Sallam (2023) [21] stated that human misuse of AI tools can lead to dependency on the tools and make people lazy;

(3) The application of AIGC technology in art and design. For example, Xu et al. (2023) [22] stated in their research that artificial intelligence has made initial attempts in art, and listed artificial intelligence generation systems such as video generation system Runway, image processing system Toolkit, social media system IO, music system Amper Music, and artistic image system Dall-E2. Cetinic & She (2022) [23] outlined the application of artificial intelligence (AI) in art analysis and creation, emphasising its importance in understanding existing artworks and generating new artworks, and looking ahead to the potential future impact of AI technology in the arts. Mazzone & Elgammal (2019) [24] explore the role of AI in artistic creation, arguing that AI-generated artworks constitute art, and advocate human-machine collaboration to expand creative potential, highlighting the contribution of AI in modelling and analysing artistic styles.

(4) Investigate the current situation of AIGC adoption intentions. Existing research has shown that various factors have a significant impact on people’s willingness to adopt AIGC technology. Gao & Xie (2023) [25] pointed out that perceived usefulness and ease of use are the two main factors influencing the adoption of AIGC technology. Their research shows that when users believe that AIGC technology can significantly improve their work efficiency and is easy to use, they are more likely to adopt the technology. Beaudry & Pinsonneault (2005) [26] also pointed out that the complexity of the technology and the learning curve are important factors influencing the willingness to adopt AIGC. Some users hesitate because they find it difficult to master AIGC technology. Meanwhile, Wang & Chen (2024) [27] explored the influence of perceived attributes and trust on designers’ behavioural intentions from the perspective of perceived risk in their study on the integration of TAMs and TPB. They found that despite the potential benefits of AIGC, designers are reluctant to adopt AIGC due to insufficient trust in the technology, especially in terms of data security and privacy. Referring to the ’gatekeeper network model’ proposed by Wu & Zhai et al. (2023), the study analysed the key influencing factors of AIGC technology adoption among designers by extending the TAM and the Technology Readiness Model. It was found that in the process of AIGC technology adoption, the technology must not only be easy to use and practical, but also effectively connect internal and external resources through the platform and the role of the "gatekeeper" [28].

In conclusion, this study identifies several gaps in the existing research on AIGC adoption intention, as evidenced by an analysis of relevant literature.

(1) The TAM model and adoption intention have not been sufficiently examined in relation to the technical and interactive characteristics of AIGC. As a novel application, AICG exhibits technical characteristics such as autonomous learning capacity, high flexibility and customisability, in addition to interactive characteristics including intelligent interaction, real-time feedback and adjustment, and multimodal interaction. However, a significant number of literature sources have not provided a comprehensive examination of the impact of AIGC characteristics on perceived ease of use, perceived usefulness, and adoption intention.

(2) The extant research concentrates on the technical functions and application effects, whereas there is comparatively little discussion of the psychological state and emotional factors of designers when adopting AIGC technology. The role of psychological factors in moderating the relationship between AIGC technology characteristics and interactive characteristics, as well as usage intentions, remains an unaddressed research gap.

(3) There is a paucity of in-depth analysis of the psychological and emotional factors of designers when using AIGC technology, particularly with regard to the research on creative autonomy, self-efficacy and technological readiness.

(4) Despite the existing literature discussing the impact of perceived usefulness and perceived ease of use on behavioural intention through the TAM model, existing research lacks a detailed analysis of the deeper influencing mechanisms behind behavioural intention. For instance, there is a dearth of systematic empirical analysis examining the direct and indirect impact of AIGC technology’s automation and personalised design capabilities on designers’ attitudes towards its use, and the subsequent translation of these attitudes into actual usage behaviour.

(5) Despite existing literature discussing the application of AIGC technology in design, there is a paucity of analysis concerning the specific effects of the technology’s technical and interactive characteristics on designers’ perceived usefulness and ease of use. The majority of studies do not provide a comprehensive analysis of the impact of these characteristics on designers’ behavioural intentions and creative processes in practice. Furthermore, the role of psychological factors, such as optimism and insecurity, in designers’ acceptance of AIGC technology has not been sufficiently addressed, particularly in the context of integrating the TAM and TRI models.

In light of these shortcomings, the present study will: (1) This study will conduct an in-depth exploration of the technical and interactive characteristics of AIGC, revealing their impact on designers’ perceived ease of use and perceived usefulness. This will contribute to the existing literature on this topic by filling the gaps in knowledge in this area. (2) It incorporates psychological factors, such as optimism and innovativeness, to analyse their impact on designers’ attitudes and behavioural intentions, thereby addressing a gap in research on designers’ psychological and emotional factors. (3) By combining the TAM and TRI models, the study presents a systematic analysis of the relationship between AIGC technical characteristics and designers’ intentions to use, thereby addressing the limitations of existing models.

Consequently, this study will integrate the aforementioned factors, including AIGC technical characteristics, interactive characteristics, perceived ease of use, perceived usefulness, and the designer’s psychology, to comprehensively examine their influence on designers’ adoption of AIGC. While addressing these gaps through empirical analysis, it also offers a novel perspective on understanding designers’ behaviour in using AIGC and provides theoretical support for the future development of AIGC.

2.2 TAM model

The TAM model, fully known as the Technology Acceptance Model, was proposed by Davis in 1989.The TAM model, with its valid and robust characteristics, is one of the most commonly used models in current research and is often used to explain users’ acceptance of new technologies and, or how to use a particular technology [29]. In the TAM model, perceived ease of use and perceived usefulness together influence whether or not a user accepts anytechnology in terms of usage imagery and attitudes, and indirectly influence actual usage behaviour [30]. In the TAM model, attitude towards use, as the third component, is an important factor used to judge an individual’s acceptance of a new technology [29]. Gao and Xie (2023) [25] state that designers are more likely to adopt AIGC technology when they perceive that the technology can improve work efficiency and is simple to use. This shows that the attitude towards use directly affects the willingness to actually use the new technology. Perceived ease of use and perceived usefulness are two factors that directly or indirectly influence the adoption attitude of users [31,32]. Where perceived ease of use shows a positive correlation for perceived usefulness, i.e., if a technology is perceived as easy to use, the degree to which it is perceived as useful is also higher [33].

Many scholars have used the TAM model in AI research [34–36], and it has been proven to be effective in predicting and explaining users’ acceptance and adoption behaviours of AI, and AIGC applications are a type of AI, so existing TAM research models and empirical methods can be applied to this study.

2.3 AIGC technology characteristics

Whether it is able to improve the efficiency of the user’s work, it has a high value of use for the user of a particular system. Tencent Research Institute (2023) [10] argues that the perceived value of AI-generated content (AIGC) to the designer community is multidimensional, and that the most basic ability of AIGC is to generate content, including text, images, video, code, 3D content, or "multi-modal content" in which several media types are exchanged and combined with each other."Jin & Ryu (2022) [30] believe that the processing of image information can be transmitted with visual information, which increases the persuasive power of imagery communication. It is due to this technical characteristic of AIGC, when designers use ai tools, it can greatly improve the design efficiency of designers. And in the design process to provide better creative combinations, especially interdisciplinary aspects, can make up for the designer in other knowledge areas, transform the designer’s design thinking [14]. It can be seen that the technical characteristics of AIGC focus on its own functionality and task execution capabilities, emphasising the performance of the system in generating content and processing tasks. The main focus is on the functionality and output results of the system itself and emphasises the performance of the system in generating content and processing tasks. Therefore, when designers adopt AIGC, the technical characteristics mainly affect perceived usefulness.

In the context of this study, the technical characteristics of AIGC are (1) AIGC’s high efficiency, which greatly improves the efficiency of designers’ content production; (2) AIGC’s freedom, which allows designers to personalise themselves using AIGC; (3) AIGC’s proactivity, which recognises and understands designers’ needs and intentions, and then automatically generates content that meets users’ requirements; (5) AIGC’srichness, which can continuously improve the quality and diversity of generated content; (5) AIGC automation, which can automatically complete the whole process of design creativity, planning to content generation.

2.4 AIGC interaction characteristics

AIGC deeply participates in a variety of creative design activities at the same time, further stimulating the effectiveness of human-computer collaboration. The interactive characteristics of AIGC are mainly reflected in its ability to effectively interact with and feedback from designers or other systems [5]. AIGC is realised through the use of AI techniques and algorithms, enabling AIGC to understand and respond to various inputs and make intelligent adjustments and feedbacks [24].AIGC interaction can better understand designers’ design needs, improve the utility of design and user experience, and at the same time provide designers with more room for innovation [15,16]. In the context of this study, AIGC interaction characteristics refer to the degree of personalised and intelligent interaction experience that the AIGC application satisfies the design needs of the designer through AI algorithms so that the designer can feel the personalised and intelligent interaction experience. It includes: (1) the interaction of AIGC applications can imitate human communication and has "social characteristics"; (2) the interaction of AIGC applications can provide a personalised experience; (3) the interaction of AIGC applications has a fast response speed; and (4) the interaction of AIGC applications has the initiative and intelligence.

The interactive features of AIGC tools significantly improve designers’ work efficiency, creative expression and overall work experience; they enable designers to focus more on creative expression and design innovation, improve work efficiency and creation quality, and give full play to their creative potential. Therefore, the interactive features more affect the designers’ operating experience and perceived ease of use, i.e., the comfort and ease of operation during the use process [37].There is a major impact on perceived ease of use when designers adopt AIGC.

2.5 Perceived interaction characteristics vs. technical characteristics

In AIGC research, technical characteristics and perceived interactivity are the two dimensions that are used to analyse the psychology of designers. Technical characteristics focus on the capability and functionality of the technology itself, paying more attention to the aspect of the capability and performance of the technology or tool in performing the task. Perceived interaction characteristics, on the other hand, generally focus on the designer’s subjective experience, i.e., the designer’s experience when interacting with the AIGC system. This characteristic emphasises the system’s ability to understand the designer’s intention quickly and accurately, and to provide feedback and adjustments in a flexible way. It mainly enhances the designer’s comfort in the process of operation and improves the efficiency of creation.

It can be seen that these two dimensions play different roles in the TAM, with the technical characteristics focusing on the functionality and output results of the system itself, and emphasising the system’s performance in generating content and handling tasks. Perceived interaction characteristics, on the other hand, mainly affect the designer’s operating experience and system ease of use, and emphasise the process of interaction between the designer and the system. Therefore, although the improvement of interaction characteristics can be a good operating experience for designers, it does not mean that new technologies and systems also have high technical characteristics. For example, a system may have a high degree of intelligent interaction capabilities, but there are technical limitations in generating high-quality content.T echnical characteristics primarily impact perceived usefulness, whereas perceived interactivity tends to more directly impact a designer’s perceived ease of use; therefore, analysing the two separately allows for a better understanding of the different ways in which the interaction experience and the technical capabilities each impact a designer’s willingness to use.

2.6 The impact of technology readiness on designers’ psychology

Technology readiness is a multidimensional psychological structure that portrays two completely opposite psychological tendencies when people are faced with new technologies, i.e., the driver and the inhibitor. The driver encourages people to use the new technology, while the inhibitor discourages people from using it, and both of them work together to determine the individual’s tendency to adopt the new technology [38,39].

Technology readiness contains four dimensions, optimism, innovativeness, discomfort and insecurity. Optimism is the individual’s maintaining an optimistic and positive attitude towards technology, i.e., the user believes that the use of new technology will increase efficiency and make work and life easier. Innovativeness is an individual’s sense of leadership towards technology, where the individual believes that he or she can master a new technology and enjoy the process of its service. Discomfort is the individual’s lack of control over the technology, perceived oppression resulting in mental anxiety and confusion, i.e., the individual perceives the technology as being too complex, the learning time becomes longer, and confidence in mastering the new technology diminishes. Insecurity is the individual’s mistrust of the technology and will have doubts about its ability to work when used. This concern often arises when an individual uses a new technology for the first time, creating doubts about what the unknown brings [38].

In the context of this study, the psychological disposition of different designers is a key factor influencing their adoption of AIGC. Therefore, this study uses technology readiness to delve deeper into the influence of designers’ psychological disposition factors in adopting AIGC. The study shows that designers’ optimism and pursuit of innovation in the face of new technologies make them more willing to embrace AIGC tools, especially in terms of enhancing work efficiency and reducing repetitive labour.

3.1 Research hypotheses

3.2 Research structure

This research builds on previous studies by integrating the TAM and the TRI with factors such as AIGC technological characteristics and interaction characteristics. In light of the aforementioned hypotheses, this study puts forth the research model depicted in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Research model.

https://doi.org/10.1371/journal.pone.0314306.g001

3.3 Definition and measurement of variables

The principal variables in this research model encompass Interactive features, AIGC Technological Characteristics, Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Attitude Toward Use (AT), Behavioral Intention (BI), Optimism (Op), Innovativeness (CS), Discomfort (Di), and Insecurity (In). The measurement indicators for these variables are presented in Table 1 for reference.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Operational definitions of variables and reference scales.

https://doi.org/10.1371/journal.pone.0314306.t001

4.1 Questionnaire design

This study used a questionnaire survey method for empirical research. Before the formal survey, a pre-test was conducted to verify the reliability and validity of the questionnaire. Between 5 January and 17 January 2024, we distributed 60 questionnaires and received 52 valid questionnaires. To improve the reliability of the questionnaire options, we analysed the reliability of the pre-test questionnaire and removed the questionnaire options that did not meet the requirements.

The formal questionnaire was distributed to designers using the AIGC app between January and February 2024. Respondents could access the questionnaire by clicking on the URL link in the questionnaire. In addition to basic personal information, respondents indicated their level of agreement or disagreement on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). During the survey, designers agreed to complete the questionnaire on a fully informed and voluntary basis, answered the research questions and had the option to withdraw at any time. A total of 513 samples were collected for this study. After removing invalid samples (e.g. logical errors and duplicates), 462 samples remained, giving a validity rate of 90.06%.

In the maximum likelihood method, the ratio of the estimated parameter to the sample size (p:n) should ideally be 1:10 [65,66]. This study has 38 test variables and 462 valid samples, with a sample size of 1:12.16, which exceeds the theoretically recommended value. Therefore, we used SPSS 23.0 and AMOS 17.0 software to analyse the collected valid data.

4.2 Scale reliability analysis

Reliability analysis, also known as consistency or stability testing, is a method of examining the stability, consistency, and reliability of measurement results. To guarantee the precision of the measurements, a reliability analysis of the valid data from the questionnaire was performed prior to the commencement of further analysis. In the field of social science research, the Cronbach’s α coefficient is a commonly employed measure for the assessment of reliability. In general, a reliability coefficient above 0.9 indicates excellent reliability, between 0.8 and 0.9 indicates very good reliability, between 0.7 and 0.8 indicates good reliability, between 0.6 and 0.7 indicates acceptable reliability, and below 0.6 suggests the need for revision.

In this study, the Cronbach’s α coefficient was employed to test the reliability of the scales using the SPSS 23.0 statistical software. As illustrated in Table 2, all reliability coefficient values exceed 0.7, thereby indicating that the data’s reliability is of an exceedingly high standard. With regard to the α coefficient if item deleted, no significant increase in reliability was observed upon the removal of any item, thereby indicating that no items should be removed. Furthermore, the corrected item-total correlation (CITC) values for all items exceeded 0.4, indicating strong correlations between the items and acceptable reliability levels. In conclusion, the data exhibits high reliability, as indicated by all reliability coefficient values exceeding 0.7, and is thus suitable for further analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Cronbach’s alpha reliability analysis.

https://doi.org/10.1371/journal.pone.0314306.t002

4.3 Descriptive statistics

The demographic characteristics of the sample are presented in Table 3. In terms of age distribution, the majority of the sample falls within the 26–35 years old category, representing 39.61% of the total. With regard to gender, the sample indicates a slight predominance of male respondents, representing 51.30% of the total, while 48.70% of the respondents are female. In terms of educational attainment, over 40% of the sample has obtained a Bachelor’s degree. A total of 20.13% of respondents identified as fashion designers.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Demographic characteristics.

https://doi.org/10.1371/journal.pone.0314306.t003

In terms of monthly income, over 20% of the sample falls within the 9,000–10,000 RMB income bracket. In terms of work experience, the most common period is that which spans one to five years, representing 36.15% of the sample. This is followed by a period of five to ten years, which represents 31.82% of the total.

The sample collected for this study is relatively evenly distributed across various demographic variables, thereby reflecting a balanced representation of the study’s target population.

4.4 Validity analysis

4.4.1 KMO and Bartlett’s test.

The construct validity of this study was analyzed using factor analysis in SPSS, which verifies whether the measurement scale meets the requirements for the research objectives. The Kaiser-Meyer-Olkin (KMO) test and Bartlett’s test of sphericity were conducted on the relevant variables using the Statistical Package for the Social Sciences (SPSS) to evaluate the suitability of factor analysis.

Validity can be defined as the extent to which a test or measurement tool accurately captures the psychological and behavioural characteristics it is designed to measure, thereby ensuring the accuracy and reliability of the results. In general, a smaller significance level for Bartlett’s test of sphericity (P < 0.05) is indicative of a higher likelihood of meaningful relationships between the original variables. The KMO value is used to assess the suitability of factor analysis by comparing the magnitude of the simple and partial correlation coefficients among items, with a value of 0 indicating no suitability and a value of 1 indicating optimal suitability. The standards for suitability for factor analysis are as follows: A KMO value greater than 0.9 indicates excellent suitability for factor analysis, while a value between 0.7 and 0.9 indicates good suitability, and a value between 0.6 and 0.7 indicates moderate suitability. A value between 0.5 and 0.6 indicates marginal suitability, while a value below 0.5 suggests that factor analysis should not be performed. Bartlett’s test of sphericity is employed to ascertain whether the correlation matrix is significantly disparate from the identity matrix. A significance value of less than 0.05 indicates that factor analysis is an appropriate methodology for the data.

As illustrated in Table 4, the KMO value for this study is 0.913, which exceeds 0.8, indicating that the data are highly suitable for factor extraction (and indirectly suggesting good validity).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. KMO and Bartlett’s test results.

https://doi.org/10.1371/journal.pone.0314306.t004

4.4.2 Exploratory factor analysis.

As can be seen in Table 5, the SPSS factor analysis shows that there are 10 factors with eigenvalues greater than 1 in the questionnaire items. The first factor explains 22.499% of the variance, which is less than the critical standard of 40%. The results show that there is no serious common method bias in the data collection of the questionnaire used in this study [67].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Explained variance table.

https://doi.org/10.1371/journal.pone.0314306.t005

Meanwhile, the cumulative variance explained is 75.262%, which indicates that the extracted common factors have high validity. As can be seen from Table 5, a total of 10 common factors were extracted, which is consistent with the previously divided dimensions, further indicating good validity.

In order to better explain the naming of each principal factor, an orthogonal rotation of the factor load matrix was performed using the maximum variance method. The size of the factor load value was used as a criterion for retention and deletion when selecting measurement items. A factor load matrix was generated and factor loads less than 0.4 were deleted. Topics in the final factor arrangement that were in the same column were placed in the same category.

As shown in Table 6, the correlation coefficients of all the research items are higher than 0.4, which means that there is a strong correlation between the research items and the factors, and the factors can effectively extract the information. After ensuring that the factors can extract most of the information of the research items, the correspondence between the factors and the research items is analysed (the absolute value of the factor load coefficient is greater than 0.4, which indicates that the item and the factor have a corresponding relationship).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Rotated factor loadings.

https://doi.org/10.1371/journal.pone.0314306.t006

4.5 Measurement model

4.5.1 Convergent validity.

We used AMOS 17.0 to analyse the structural equation model. As a large number of studies use AMOS for analysis, AMOS has proven to be a reliable software for structural equation modelling. Anderson & Gerbing (1988) [68] suggest that data analysis can be divided into two stages. The first stage is the measurement model, which uses maximum likelihood estimation to estimate parameters such as factor loadings, reliability, convergent validity and discriminant validity. The second stage is a convergent validity study based on Hair et al. (1998) [69], Raykov& Marcoulides (2011) [70], Fornell & Larcker (1981) [71] and Chin (1998) [72].

As shown in Table 7. The itemised standardised factor loadings for all variables are greater than 0.7, indicating that each observed variable contributes significantly to the explanation of its latent variable. It should be noted that the CR is greater than 0.8, which is significantly higher than the standard of 0.7. This indicates that the observed variables under each dimension can explain the dimension well. The AVE values in this study are all above the standard value of 0.6, indicating that the convergent validity of the scale is good.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Summary of determinants analysis.

https://doi.org/10.1371/journal.pone.0314306.t007

4.5.2. Discriminant validity.

The assessment of discriminant validity was conducted in accordance with the approach proposed by Fornell & Larcker (1981) [71]. The discriminant validity of a construct is established when the square root of the average variance extracted (AVE) for that construct is greater than the absolute value of the correlations between that construct and the other constructs in the model. This suggests that the constructs in the model are distinct from one another.

The results demonstrate that the diagonal values (representing the square roots of the AVEs) exceed the off-diagonal correlation values, indicating that all constructs in this study possess satisfactory discriminant validity (see Table 8). It can therefore be concluded that each dimension in this study exhibits high discriminant validity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Discriminant validity of measurement model.

https://doi.org/10.1371/journal.pone.0314306.t008

4.6 Structural model analysis

4.6.1 Model fit indices.

A model with a good fit indicates that the model closely reflects the sample. In accordance with the recommendations of Jackson et al. (2009) [66], Kline (2015) [73], Whittaker (2011) [74], Hu & Bentler (1999) [75], and other scholars, this study employed a range of fit indices to assess the fit of the structural model. The selected fit indices are as follows: CMIN, degrees of freedom (DF), CMIN/DF, Normed Fit Index (NFI), Tucker-Lewis Index (TLI), Comparative Fit Index (CFI), Goodness of Fit Index (GFI), Relative Fit Index (RFI), and Incremental Fit Index (IFI).As illustrated in Table 9, the majority of the model fit indices align with the optimal standards, substantiating the model’s commendable fit.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 9. Model fit indices.

https://doi.org/10.1371/journal.pone.0314306.t009

4.6.2 Path analysis.

As evidenced in Table 10, when the interactive and technical features influence perceived ease of use, the standardized path coefficients are all greater than 0, and p is less than 0.01. This indicates that the interactive and technical features have a significant positive impact on perceived ease of use.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 10. Path coefficients.

https://doi.org/10.1371/journal.pone.0314306.t010

When technology features and perceived ease of use affect perceived usefulness, the standardized path coefficients are all greater than 0 and p is less than 0.01, indicating that technology features and perceived ease of use have a significant positive impact on perceived usefulness. When perceived usefulness and perceived ease of use affect usage attitude, the standardized path coefficients are all greater than 0 and p is less than 0.01, indicating that perceived usefulness and perceived ease of use have a significant positive impact on usage attitude. When perceived usefulness, optimism, innovative spirit and attitude towards use affect behavioural intention, the standardized path coefficients are all greater than 0 and p is less than 0.01, indicating that perceived usefulness, optimism, innovative spirit and attitude towards use have a significant positive impact on behavioural intention. In the case of a negative correlation between behavioural intention and insecurity, the standardised path coefficient is less than 0 and p is less than 0.01, indicating that insecurity exerts a significant negative influence on behavioural intention. The effect of discomfort on behavioural intention is not significant at the 0.05 level, indicating that discomfort has no impact on behavioural intention. When optimism and innovative spirit affect the technical characteristics, the standardized path coefficients are all greater than 0 and p is less than 0.01, indicating that optimism and innovative spirit have a significant positive impact on the technical characteristics. In the case of a significant correlation between insecurity and discomfort and the technical characteristics, the p-value is not significant when greater than 0.05. This indicates that there is no impact of insecurity and discomfort on the technical characteristics.

4.7 Hypothesis interpretation

As illustrated in Fig 2, the regression coefficients of the structural equation model (SEM) employed in this study are presented. A higher coefficient indicates that the independent variable exerts a more pronounced influence on the dependent variable.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Hypothetical path coefficients diagram.

https://doi.org/10.1371/journal.pone.0314306.g002

The hypotheses H1b, H2a, H2b, H3a, H3b, H4a, H4b, H5, H6a, H6b, H7a, and H7b are supported by the model, indicating that the relationships between variables were found to be significant.

The hypotheses H1a, H8a, H8b, and H9a are not supported, indicating that these proposed relationships did not have a significant effect on the model.