Sage Journals: Discover world-class research

Abstract

As artificial intelligence continues to evolve, its application in the realm of design through Artificial Intelligence Generated Content (AIGC) has seen a marked increase. However, limited research exists on the effects of technostress and task-technology fit on designers’ sustained usage intentions toward AIGC. This study aims to identify key factors influencing designers’ continuous intention to use AIGC. It integrates the Unified Theory of Acceptance and Use of Technology (UTAUT) with the Task-Technology Fit (TTF) model and introduces technostress as a moderating variable. A theoretical framework was constructed to explain the continuous intention of using AIGC among designers. The research utilized a stratified purposive sampling method to enlist participants, gathering data from 443 design students and professional designers in China. Data were analyzed using structural equation modeling. The results reveal that: (a) designers’ performance expectancy from AIGC has a strong positive effect on continuous intention; (b) both technology and task characteristics substantially enhance task-technology fit, which in turn indirectly augments continued intention via user satisfaction; (c) an interaction between UTAUT and TTF models enhances performance expectancy through task-technology fit; (d) technostress negatively moderates the impact of performance and effort expectancy, along with social influence and user satisfaction, on continuous intention. These results provide actionable insights for the design industry to facilitate the effective and durable adoption of AIGC. Additionally, the study offers strategic recommendations for AIGC developers and managers to optimize user experience and sustain user engagement.

Plain Language Summary

Understanding What Drives Designers to Keep Using AIGC Tools in Their Work

Artificial Intelligence (AI) is becoming more common in the design world, especially through tools that generate content automatically, known as AI-generated content (AIGC). While these tools can be powerful, it is important to understand what makes designers want to keep using them over time. This study explores the main reasons why designers choose to continue using AIGC tools, focusing on how well these tools match the designers’ tasks and whether they cause stress (called technostress). We surveyed 443 design students and professionals in China to find out more. The results show that designers are more likely to keep using AIGC tools if they believe these tools will help them do a better job. Additionally, if the tools are well-suited to the specific tasks designers need to complete, they are more satisfied and more likely to stick with them. However, if the tools cause too much stress, it can reduce designers’ willingness to keep using them. These findings suggest ways for the design industry to improve AI tools and make them more user-friendly, helping designers to use them effectively and for a longer period.

Keywords

AIGC UTAUT TTF continuous intention technostress

Introduction

The escalating fascination with Artificial Intelligence Generated Content (AIGC) is poised to reshape content creation workflows significantly. Utilizing a blend of AI, NLP, and computer vision, AIGC can autonomously produce textual, visual, and auditory content, as detailed by Lin et al. (2023). This technology not only creates and modifies design content but also enhances design languages and expressions, potentially catalyzing transformative impacts within the design industry (Hou & Xu, 2021). AIGC primarily aims to streamline the content production process, offering tailored content creation services to individuals (Cao et al., 2023). In conclusion, AIGC is in a high demand and has many purposes and benefits, including creating individualized content quickly and its use in different media and fields.

Many benefits will accrue to design professionals as a result of AIGC. An empirical study on large-scale information and machine-made product planning also shows that the major changes in design occur when the system is guided by AI combined with computers (Quan et al., 2023). This is in agreement with another study that has identified key technological tools in this field and also stressed the significance of AI in design planning (Sein et al., 2011). Prior literature on intelligence amplification shed light on how AI applications are pivotal in the design process, especially in its early stages, to facilitate and organize communication and ideas with the stakeholders (Sebastian Piest et al., 2022). AI applications have rapidly developed in environmental design, profoundly transforming design tools and methods, and enriching design content, further supporting the potential of AIGC in the design industry (Hou & Xu, 2021). Additionally, the use of AI in cultural, creative, and design industries has been described as transformative, indicating how it changes production and operation methods in these industries (Zhao & Zhang, 2023).

Current research related to AIGC is oriented towards three major directions: (i) the theoretical feasibility and technical advantages of AIGC applied to the design industry. Yang et al. (2021) demonstrated the efficiency and feasibility of intelligent product form design based on design cognitive dynamics and a cobweb structure, providing practical evidence supporting the viability of AIGC in product design. Jian et al. (2021) analyzed the impacts of AI on teaching design, offering insights into the paradigm shift and connotation analysis of teaching design under AI technology; (ii) the potential risks and shortcomings associated with AIGC. For example, safety issues of AIGC were not included in standard design practices, raising questions about whether AIGC could address these problems in the construction industry (Gambatese et al., 2005). Furthermore, the rapid progress of AIGC has raised significant concerns about service latency, security, and trustworthiness, which are crucial in the design industry (Y. Liu et al., 2023). Moreover, despite the increasing need for new technologies and materials, industry reports reveal that most design firms retain legacy workflows due to institutional inertia and high retraining costs (Statista, 2023). This stagnation in methodological evolution attributed to resistance to disruptive technologies and fragmented AIGC adoption (Xu et al., 2023) may also be an indication of AIGC’s limitations.

Menon et al. (2023) proposed the study based on the Unified Theory of Acceptance and Use of Technology (UTAUT) model to investigate the factors that affect the acceptance of ChatGPT. P. Zhou et al. (2023) focused on the impact of anthropomorphism in AI voice assistants on the consumer’s cognitive responses and their long-term usage intentions. In a related study, Jo (2023) explored the factors affecting the users’ behaviors toward AI applications such as ChatGPT, and observed significant differences in the roles of personalization-satisfaction, perceived utilitarian value of the application and word of mouth (WOM) recommendation, as well as the relationship between behavioral intentions and WOM among students and professionals. Nonetheless, the scope of research focusing on designers’ propensity to engage with AIGC remains sparse. While AIGC’s technical feasibility is well-documented (Cao et al., 2023; Yang et al., 2021), scant research examines its sustained adoption among designers—a gap highlighted by recent critiques (Li, 2024). Current studies predominantly focus on risks like output inconsistency (Gambatese et al., 2005) or ethical concerns (Zhou et al., 2023), neglecting the psychological and task-contextual barriers explored here.

This study aims to construct a comprehensive model to interpret designers’ intentions for the continuous use of AIGC. The term “comprehensive” here refers to the integration of two established theoretical frameworks—the Unified Theory of Acceptance and Use of Technology (UTAUT) and the Task-Technology Fit (TTF) model—while introducing technostress as a novel moderating variable. By combining UTAUT (which focuses on users’ perceptions of technology utility and social influence) with TTF (which emphasizes the alignment between technology capabilities and task requirements), our model addresses both the psychological drivers of adoption (e.g., performance expectancy, effort expectancy) and the contextual fit of technology to specific design tasks. Furthermore, the inclusion of technostress as a boundary condition extends prior research by accounting for the psychological barriers that may counteract the benefits of AIGC adoption. This multi-dimensional approach enables a holistic understanding of designers’ sustained engagement with AIGC, bridging gaps between technology acceptance theories and task-environment dynamics.

While many designers are attracted to AIGC, a significant number have discontinued its learning and use. Therefore, it is essential to explore the factors influencing designers’ continuous intention to use AIGC. We consider technostress as a boundary factor, which, combined with the UTAUT and TTF models, helps to more comprehensively understand and predict the factors and mechanisms that determine designers’ acceptance and continuous intention to use AIGC.

The notion of continuous intention is typically associated with the propensity to persist in using information technology systems, as explained by Venkatesh et al. (2003).reconciled eight theoretical models, including TAM, TPB, and IDT, to build a new theoretical model—the UTAUT. While UTAUT demonstrated superior explanatory power in users’ continuous intention for information systems, accounting for seventy percent of users’ intentions, it has its limitations (Venkatesh et al., 2003). For instance, even if designers find AIGC useful and easy to use, they may still choose not to utilize it if it does not fit the task requirements of designers and design students. Therefore, the TTF model is also introduced in this study.

The TTF model asserts that the adoption and sustained use of technology are primarily influenced by how well the technology’s features correspond to the users’ task demands (Goodhue, 1995). Numerous studies integrating the UTAUT with the TTF model have examined the adoption of informational services, including e-learning in healthcare, smart home healthcare services, and mobile banking systems (Alkhwaldi et al., 2022; Kang et al., 2022; Oliveira et al., 2014). These investigations reveal that the combination of factors from both models significantly affects users’ decisions to adopt technologies. Consequently, this research aims to create an enhanced conceptual framework by merging the UTAUT and TTF models to explore the determinants and mechanisms influencing designers’ intentions to utilize AIGC.

Technostress, as identified by Brod (1984), is an adjustment disorder stemming from the challenges individuals face when attempting to adapt to new technologies in a manner that promotes well-being. This has become worse for employees today because of the dynamism in the advancements of Information and Communication Technologies (ICTs). Technology stress which is defined as the pressure that ICT users face when trying to learn and operate with new technology has been widely investigated (Ragu-Nathan et al., 2008). This stress includes physical, mental, and behavioral reactions to present and future technological requirements (Hagemann & Klug, 2022). Also, the use of digital tools lead to stress which negatively impacts job satisfaction (Suh & Lee, 2017). In conclusion, technostress is a serious problem due to fears from different sources, technology, its correct utilization, or the shift in technology and advancements. The role of technostress as a barrier to adopting and effectively employing novel technologies is increasingly recognized. Research such as that by Yener et al. (2020) has illustrated that factors like technological self-efficacy and effective time management can influence technostress, subsequently impacting employee performance with burnout serving as a mediating variable. Additionally, Zhang et al. (2022) investigated how technostress might mediate the propensity for knowledge concealment among R&D staff in the era of digitalization. Based on these insights, this paper posits technostress as a moderating variable and assesses its potential moderating influence.

While existing studies have explored initial adoption factors of AI tools in creative industries, there remains a critical gap in understanding the mechanisms driving designers’ long-term commitment to AIGC systems. Prior research has predominantly focused on technical capabilities and immediate usability metrics, overlooking two pivotal dimensions: the congruence between technological features and design-specific task requirements, as well as the psychological barriers induced by rapid technological evolution. The current literature provides limited insights into how the interplay of task-technology alignment and technology-induced stress jointly shape sustained usage patterns in professional design contexts. This study addresses these underexplored areas by developing an integrated theoretical framework that bridges technological acceptance theories with human-computer interaction stressors. Specifically, we investigate how performance expectations shaped by task-technology fit interact with technostress to influence designers’ continuous engagement, offering novel perspectives on sustaining AI adoption in creative workflows.To systematically investigate these dynamics, we formulate the following research questions:

RQ1: How does the UTAUT influence designers’ continuous intention of using AIGC?

RQ2: How does the TTF influence designers’ continuous intention using of AIGC?

RQ3: Whether technostress plays a moderate role?

Subsequently, an examination of prior studies focusing on AIGC, technostress, and the sustained usage intentions of designers regarding AIGC will be conducted. This will be followed by an elaboration of the research methodology employed, the analysis of the data collected, and the outcomes achieved. A comprehensive discussion on these results will conclude this section. The final part of the article will summarize our research findings and propose future research areas. This research focuses on designers’ continuous intention of using AIGC, providing a conceptual framework and practical recommendations for its further application in the design industry. Additionally, this research incorporates technostress to investigate designers’ acceptance and apprehension of new technologies. Within the UTAUT and TTF models, we examine how technostress and task-technology fit work together to influence designers’ continuous intention of using AIGC. Understanding these relationships helps AIGC developers and business managers strategically improve user experience and promote continuous intention of using AIGC.

Theoretical Background and Hypotheses

AIGC

AIGC employs AI technology for content creation based on user requirements. This technology shows significant promise with numerous applications (Ballardini et al., 2019). Initially, AIGC primarily assisted in content generation using fixed templates, particularly in professional scenarios such as video production, entertainment, and modeling. It has since evolved into a cutting-edge technology that revolutionizes various industries through the automation of content creation with advanced machine learning algorithms. A pivotal moment occurred in 2014 when Goodfellow et al. (2020) proposed Generative Adversarial Networks (GANs), which use adversarial learning to generate data. In 2015, Alexander Mordvintsev discovered that neural networks could be trained in reverse, enabling the generation of synthetic graphics and the recognition of real ones. This led to the creation of Deep Dream, a psychedelic computer vision program that significantly impacted the art world (Miller, 2019). Mahajan et al. (2018) achieved accelerated training of models, such as VirTex, ICMLM, and ConVIRT, using accelerators on millions to billions of images, reducing training time to days for 1 to 200,000 images. In 2021, Radford et al. (2021) developed the CLIP (Contrastive Language-Image Pre-Training) algorithm, which excels in learning visual features and enabling multi-modal pre-training. Nichol and Dhariwal (2021) demonstrated that diffusion models, which are likelihood-based with stationary training objectives, surpass GANs in sample quality. Furthermore, Ho et al. (2020) developed a diffusion model for graph generation using forward diffusion and reverse generation processes. Following these advancements, AIGC applications have expanded into digital twins, digital modeling, and artistic creation.

AIGC is widely used in the design industry, encompassing advertising, animation, graphics, web, and product design. Designers have incorporated AI tools and algorithms to enhance creativity, expedite workflows, and generate new content. AIGC technology has transformed the creation process by enabling the efficient production of high-quality content. It opens opportunities for designing intelligent buildings, environmental design, and arts and crafts. AI in design extends beyond traditional fields and includes applications in Chinese handicraft designs (Xiang et al., 2022). The future of AI in design is promising, with the advent of AI-driven information ecosystems and AI-assisted design systems (Guo, 2020). Research into the potential challenges and future developments of AI in design also highlights the importance of AI-assisted physical layer design in wireless systems, which is expected to emerge soon, and the integration of AI with art and design (Hu et al., 2020).

AIGC is regarded as a popular creative method among designers. They employ textual keywords and models to generate design outcomes. AIGC empowers designers to explore diverse design concepts from novel perspectives. By fostering serendipitous discoveries, these tools expand creative capabilities beyond conventional means. These tools facilitate data manipulation and provide innovative solutions to challenging issues. AIGC tools such as DALL-E 2, Midjourney, and Stable Diffusion are increasingly employed in design disciplines. Developed by OpenAI, DALL-E 2 merges natural language processing with computer vision technologies to transform textual descriptions into visual outputs (Radford et al., 2021). This capability makes it particularly valuable for applications spanning from design and illustration to broader artistic endeavors (Hanafy, 2023). Another commonly used AIGC tool is Midjourney. In this tool, users input content, which is then converted into the corresponding image by an AI program. This tool serves multiple purposes, helping artists and creators quickly visualize their ideas and assisting those with reading difficulties by visually representing written content (V. Liu & Chilton, 2022). Stable Diffusion, a free tool initiated in 2022 by Stability AI, CompVis, and Runway, is available for public use. Following the concept of DALL-E 2, Stable Diffusion is a robust tool for modeling stable structures and mixing processes. It is widely used in various industries and research centers for designing and visualizing objects such as buildings or bridges without actual construction (Al-Hammadi et al., 2022). Additionally, this software promotes environmentally sustainable practices, aiding professionals in business and academia in developing long-term sustainable designs (Oyedeji et al., 2018). It has also been employed to study stable diffusion and image classification through brain functions, demonstrating its versatility across different fields (Raglin & Basak, 2023).

Despite the availability of numerous AIGC products developed for the design industry, many users are dissatisfied with the resulting designs. Although AIGC can deliver efficient results, a significant gap between the output and the designer’s expected results greatly reduces user satisfaction, thereby diminishing the intention to continue using AIGC. For some designers, the appeal lies in the simplicity of operation, low learning cost, and user-friendliness of AIGC. Conversely, other designers perceive it as conflicting with their conventional design logic and creative process, leading to their reluctance to adopt AIGC. Related research has demonstrated that the diffusion and widespread application of new technology largely depend on users’ perceived usefulness. Typically, the initial engagement with AIGC by designers and design students is driven by this perceived usefulness. The principal impetus for the advancement of AIGC resides in the sustained interest of its users. Numerous designers and students within the design discipline are drawn to the novel, creative potentials facilitated by AIGC. Nonetheless, the continuation of utilizing this technology hinges not only on personal or contextual influences but also on the extent to which the technology and the specific tasks are congruent in delivering the anticipated results.

UTAUT Model

Introduced by Venkatesh et al. (2003), the UTAUT provides a comprehensive framework for assessing and predicting the acceptance and utilization of innovative technologies. The framework posits that four core constructs—performance expectancy, effort expectancy, social influence, and facilitating conditions—critically influence an individual’s intent to adopt a technology. Extensively applied in research on technology adoption, the UTAUT model has facilitated numerous studies. For instance, in 2021, Cabrera-Sánchez et al. (2021) utilized this model to investigate the factors influencing Spanish consumers’ adoption of artificial intelligence applications, highlighting significant determinants of technology acceptance. Their results indicated that the variables proposed in the UTAUT model significantly influenced five different behavioral aspects, moderated by techno-fear and consumer trust. Du et al. (2023) proposed an extended UTAUT model highlighting three additional factors: design focus, AI anxiety, and AI quality. Despite the UTAUT model’s popularity, limited research has used it to test the continuous intention of using AIGC.

Following Venkatesh et al. (2003), continuous intention is defined as the sustained willingness to use a technology post-adoption, distinct from initial intention, which captures pre-adoption expectations. This distinction is critical, as discontinuation often arises from post-adoption dissonance between expectations and reality (Davis & Venkatesh, 1996).

Performance expectancy in this study is defined as the extent to which designers anticipate that using AIGC will improve their design efficiency. As AIGC enhances this efficiency, it is anticipated that designers’ motivation to persist with its use will increase proportionally. Effort expectancy is related to the amount of effort required by designers to master and employ AIGC effectively. It is hypothesized that when AIGC is perceived as user-friendly, simple to learn, and easy to access, there will be a greater propensity among designers to continue its use. Social influence encompasses the effects that peers, mentors, and colleagues have on a designer’s decision to keep using AIGC, as their opinions and attitudes can significantly sway such decisions. Xu et al. (2023) observed that AIGC is still in its early stages in China, facing widespread misuse and quality issues. Its application is often unstable and unsustainable. Many commercial AIGC tools, while providing easy access, fail to offer a satisfactory user experience, resulting in the loss of professional designers. Due to these factors, facilitating conditions proved challenging to assess accurately in this study. Consequently, this research excludes facilitating conditions. Following the UTAUT framework, facilitating conditions (e.g., technical support, infrastructure) are critical for technology adoption. However, in the context of AIGC adoption among Chinese designers, two key considerations led to the exclusion of this construct:

Early-stage technology maturity: AIGC tools in China are predominantly in a nascent phase, with limited institutional or organizational support systems (e.g., standardized technical assistance, training programs). As reported by Xu et al. (2023), most commercial AIGC platforms in China currently lack stable technical infrastructure, making facilitating conditions inconsistent across users.

Focus on individual-level drivers: This study prioritizes individual perceptions (e.g., performance expectancy, effort expectancy) and task-context alignment (TTF) as primary determinants of continuous intention. Including facilitating conditions would shift the focus to organizational or environmental factors, which are beyond the scope of this research.

Future studies may incorporate facilitating conditions once AIGC ecosystems in China mature and standardized support mechanisms emerge.

Our study acknowledges that individual differences also impact the relationships in the UTAUT model. While UTAUT traditionally includes age and gender as moderators, their exclusion in this study is justified by two factors. First, the integrated model’s complexity (nine latent variables and thirteen hypotheses) prioritizes parsimony to avoid overfitting (Hair et al., 2006). Second, prior meta-analyses (Venkatesh et al., 2003) indicate that demographic variables often yield weaker explanatory power compared to core constructs like performance expectancy. Additionally, our sample exhibited homogeneity in age (64.56% aged 21–30) and gender (55.76% female), reducing their potential moderating effects (Table 1). Future studies could explore these variables in extended models with larger, diversified samples (Figure 1).

Table 1.

Statistics of Respondents.

Category	Option	Frequency	%
Valid samples (N = 443)
Gender	Male	196	44.24
Gender	Female	247	55.76
Age	<20	58	13.09
	21–30	286	64.56
	31–40	79	17.83
	>40	19	4.29
Occupation	Designer	169	38.15
Occupation	Design student	274	61.85
Major	Graphic design/visual communication	116	26.19
	Industrial design/product design	179	40.41
	Interaction design/digital media design	74	16.70
	Public art design	25	5.64
	Interior design/display design	23	5.19
	Other design industries	26	5.87

Figure 1.

The hypothetical model.

Accordingly, the study formulates these hypotheses:

Hypothesis 1 (H1): Performance expectancy has a positive effect on the continuous intention of using AIGC.

Hypothesis 2 (H2): Effort expectancy has a positive effect on the continuous intention of using AIGC.

Hypothesis 3 (H3): Social influence has a positive effect on the continuous intention of using AIGC.

TTF Model

The TTF model suggests that technology can enhance job performance significantly when aligned with the specific demands of a task (Goodhue & Thompson, 1995). Its core assertion holds that technology yields optimal results when tailored to the precise needs of a job. On the other hand, discrepancies between technology capabilities and job requirements can lead to inefficiencies. The TTF model delineates three pivotal elements that determine technology adoption: characteristics of the task, attributes of the technology, and their alignment, termed task-technology fit. Renowned within the domain of information systems, this model serves as an essential theoretical framework for assessing and selecting information systems. In their study, Wan et al. (2020) explored the determinants of college students’ ongoing engagement with MOOCs and revealed an indirect influence of TTF on sustained usage intentions via satisfaction, attributing this to the interplay between user satisfaction, performance expectancy, and TTF. Similarly, Zhou et al. (2010) merged TTF with the UTAUT model to devise a conceptual framework for mobile banking adoption, pinpointing performance expectancy, TTF, and social influence as key drivers of technology uptake. Despite its widespread application across various research domains, TTF’s utilization in the field of AIGC remains limited. Our study seeks to leverage the TTF model to augment the thoroughness of our research.

Based on the TTF model, a favorable task-technology fit is expected to increase designers’ acceptance of AIGC and contribute to perceived satisfaction. For instance, even if AIGC boasts technical appeal or functional advantages, designers may still exhibit dissatisfaction and discontinue use if they perceive a misalignment with their task requirements or find the desired task difficult to achieve. We hypothesize that task-technology fit affects designers’ perceived satisfaction with AIGC and indirectly affects their continuous intention. Additionally, task-technology fit may positively impact designers’ performance expectancy, as a superior task-technology fit is likely to enhance designers’ performance expectancy for AIGC.

Therefore, we list the following research hypotheses:

Hypothesis 4 (H4): Task characteristics have a positive effect on task-technology fit.

Hypothesis 5 (H5): Technology characteristics have a positive effect on task-technology fit.

Hypothesis 6 (H6): Task-technology fit has a positive effect on performance expectancy.

User Satisfaction

User satisfaction encompasses individuals’ evaluations of a system or service, focusing on its usefulness, and effectiveness in achieving its designated goals (Lee, 2013). This satisfaction is a multifaceted concept influenced by factors such as perceived ease of use, perceived usefulness, and perceived control (Amalia & Fahrudi, 2021). It reflects the attitudes and perceptions individuals hold toward specific systems or products (Phan et al., 2016). Although direct assessments of user satisfaction are absent from the UTAUT model, they are derivable through performance expectations and the effort required to utilize the technology. Research by various scholars has also highlighted the significance of user satisfaction in the continued usage of diverse technologies. For example, Ashfaq et al. (2020) observed that satisfaction with a Chatbot’s support positively influences the ongoing trust users place in the Chatbots. In a similar vein, Roca et al. (2006) devised an e-learning acceptance model and observed that ongoing usage depends on satisfaction with the learning approach. It is crucial to acknowledge the influence of user satisfaction on designers’ ongoing engagement with AIGC. From these observations, the following hypotheses are posited:

Hypothesis 7 (H7): Task-technology fit has a positive effect on user satisfaction.

Hypothesis 8 (H8): Effort expectancy has a positive effect on user satisfaction.

Hypothesis 9 (H9): User satisfaction has a positive effect on designers’ continuous intention.

Technostress

Introduced by Brod (1984), the concept of technostress refers to the anxiety and strain individuals experience when adapting to new technological environments. It highlights the negative impact of technology on both the personal and professional aspects of life. Defined by Tarafdar et al. (2017), technostress manifests as psychological and emotional distress caused by technological demands at work. These elements not only jeopardize mental well-being, leading to disorders such as anxiety and depression, but they also contribute to physical health issues, including sleep disruption, and chronic fatigue (Galluch et al., 2015). Further studies by Tarafdar et al. (2010) suggest that technostress can diminish satisfaction with ICT and adversely affect the decisions regarding the adoption and continued use of technology (Chen et al., 2019; Joo et al., 2016). Hsiao (2017) supports the view that high technostress negatively influences the continued use of technology. According to Verkijika (2019), an elevated stress level from technology use inversely affects the perceived value of digital books and users’ intentions to utilize them. In essence, heightened technostress reduces the perceived benefits of online resources, thereby decreasing the likelihood of their adoption. The complex nature of technostress and its implications for long-term technology usage remains underexplored, highlighting the need for further investigation into how technostress and related factors such as fatigue influence the cessation of technology use (Maier et al., 2015).

Considering the potential impact of technostress from AIGC as a moderating factor influencing designers’ continuous intention to use AIGC, two opposing scenarios are hypothesized: Thus, if designers are able to control their technostress it may minimize or eliminate their usage of AIGC. However, if they were to effectively manage technostress, their ongoing intention to use AIGC would be boosted.

Therefore, we propose the following research hypotheses:

Hypothesis 10 (H10): Technostress negatively moderates the relationship between performance expectancy and continuous intention.

Hypothesis 11 (H11): Technostress negatively moderates the relationship between effort expectancy and continuous intention.

Hypothesis 12 (H12): Technostress negatively moderates the relationship between social influence and continuous intention.

Hypothesis 13 (H13): Technostress negatively moderates the relationship between user satisfaction and continuous intention.

Critical Perspectives on Technology Adoption Frameworks

While the Unified Theory of Acceptance and Use of Technology (UTAUT) and Task-Technology Fit (TTF) models offer robust frameworks for analyzing technology adoption, emerging scholarship has raised critical concerns regarding their applicability in creative domains such as design. Firstly, the utilitarian orientation of UTAUT’s performance expectancy construct (Venkateshet al., 2003) fails to adequately encapsulate designers’ aesthetic evaluation processes. Empirical evidence from architectural design studies (Mazzone & Elgammal, 2019) demonstrates that creative professionals prioritize expressive potential over efficiency metrics—a dimension conspicuously absent in conventional adoption models.

Secondly, the TTF model’s presumption of static task-technology alignment (Goodhue & Thompson, 1995) contradicts the iterative essence of design workflows. The data shows that some designers make multiple revisions when adjusting the output of artificial intelligence-generated content (AIGC; Brynjolfsson & McAfee, 2017), thereby suggesting dynamic rather than fixed task-technology relationships.

Thirdly, existing models predominantly disregard power dynamics inherent in organizational technology adoption. Within design firms, junior designers’ utilization of AIGC frequently reflects managerial directives rather than autonomous choice (Yin et al., 2023), thereby contravening UTAUT’s foundational assumption of voluntarism. Such institutional coercions may artificially elevate continuance intention metrics while obscuring underlying resistance.

Materials and Methods

Participants and Data Collection

In the initial part of the questionnaire, we included a question to determine eligibility: “Have you currently used AIGC?” This question helped exclude those who had not engaged with AIGC. Respondents who answered “no” were not included in the survey tally. The primary mode of the survey was digital, complemented by printed questionnaires. Digital surveys were disseminated through social media platforms, including QQ and WeChat, whereas their article counterparts were handed out in the classrooms of students. Mueller (1997) posited that a sample size should ideally range from 10 to 15 times the quantity of observed variables. For the 31 variables in the present study, the sample size was determined to be between 310 and 465 to maintain statistical robustness.

To ensure the ethical integrity of our research, the study was approved by the ethics committee of the authors’ institution. The study design minimized potential risks of harm to participants, as it only involved an anonymous, non-invasive survey. The potential benefits of this research for both the design industry and AIGC developers are significant, outweighing the negligible risks to participants. All participants were furnished with detailed information before gathering data and their consent was obtained in writing. They were assured that they could exit the study at any point without facing any negative consequences, in alignment with the recommendations of the ethics committee.

Participants were recruited through stratified purposive sampling. First, the target population was divided into two strata: (1) design students from universities across China and (2) professional designers from multiple provinces. Questionnaires were distributed via university instructors and design firms, leveraging their access to relevant populations. To ensure relevance, only respondents who confirmed prior use of AIGC tools (via the screening question, “Have you currently used AIGC?”) were included in the final sample. The majority of respondents in our survey were aged between 21 and 40. Survey data from September 2023 indicated that approximately 68% of the early adopters of AIGC applications in China were born in the 1980s to 1990s (Statista, 2023). Therefore, the age distribution of our sample aligns well with the target demographic.

Instruments

The survey is structured into two distinct sections. The first section is dedicated to collecting demographic information from participants, including their gender, age, occupation, area of academic focus, and prior engagement with AIGC technologies. The subsequent section involves a series of items crafted to evaluate the core aspects of the study’s theoretical framework. A structural model underpins the design of the questionnaire, featuring nine measurement constructs, each encompassed by 31 items. These constructs are as follows: performance expectancy, which includes three items; effort expectancy, which comprises four items; social influence and task characteristics, each detailed by four items; technology characteristics, task-technology fit, and continuous intention, each described through three items; along with user satisfaction and technostress, each delineated by four and three items respectively. For additional details, the Appendix is available for consultation. The items for this part of the survey have been adapted from existing surveys that leverage the UTAUT model, as noted in the work of Escobar-Rodríguez et al. (2014), the TTF model as documented by W. T. Lin and Wang (2012), and the synthesis of UTAUT and TTF models in the studies by Cai et al. (2023), Oliveira et al. (2014), Wan et al. (2020), as well as technostress, explored by Pirkkalainen et al. (2019). A 5-point Likert scale, from 1, indicating “strongly disagree,” to 5, indicating “strongly agree,” is utilized for capturing responses in this segment.

Data Analysis

Data analysis was performed using SPSS version 27 and AMOS version 24. The reliability of the dataset was evaluated through the computation of Cronbach’s alpha coefficient with SPSS, while statistical descriptions were also generated. These descriptions revealed an average age range of 21 to 35 years among the 443 participants surveyed. A single-sample t-test, with results showing t(442)4= .029, p 29.977, indicated that the mean age of participants aligns with the average age of the population, suggesting an absence of age-related non-response bias. The cohort consisted of 196 males and 247 females. The Chi-square test, χ² (1, N 1,443)4= 5.87, p = .15, confirmed that the sample’s gender distribution mirrors that of the general population, implying no gender-based non-response bias. To address potential Common Method Bias (CMB), the Harman single factor test was applied, revealing that the largest factor explained only 33% of the total variance, well below the 50% threshold, which indicates a minimal influence of CMB on the data. In addition to the Harman single factor test, we further assessed the potential for common method bias by employing the CFA-based unmeasured latent variable (UMLV) technique, as recommended by Podsakoff et al. (2023) . This approach involved a three-model comparison: (A) a baseline CFA model (Trait-only Model 1), (B) a Harman’s single-factor model (Method-only Model 2), and (C) an UMLV model (Method-and-Trait Model 3), introducing a single factor that links to all the indicators of the substantive correlated constructs. Further analytical procedures were carried out using AMOS, including confirmatory factor analysis to ascertain both reliability and validity of the measurement model and structural equation modeling to examine the research hypotheses.

Results

The Measurement Model

Before conducting measurement and structural model analyses, a normality test was performed for each variable. Data are considered normally distributed when the values of kurtosis and skewness fall between |3| and |10|, respectively (Moorthy et al., 2019). The results, as presented in Table 2, indicate that the kurtosis values range from −1.307 to 3.627, and skewness values from −.885 to .325. Based on these findings, the data are determined to be adequately normally distributed.

Table 2.

Results of Construct Validity, Reliability Analysis, and Normality Testing.

Construct	Item	Mean	SD	Factor loadings	CR	AVE	Alpha	Skewness	Kurtosis
Performance expectancy (PE)	PE1	4.03	.82	.846	.85	.66	.852	−1.011	.875
	PE2			.749				−.735	.354
	PE3			.838				−1.055	1.077
Effort expectancy (EE)	EE1	3.44	.79	.698	.83	.56	.832	−.239	−.490
	EE2			.660				−.349	−.277
	EE3			.791				−.345	−.288
	EE4			.829				−.307	−.251
Social influence (SI)	SI1	3.90	.71	.642	.78	.47	.783	−.774	.568
	SI2			.735				−.654	−.069
	SI3			.645				−.833	.625
	SI4			.727				−.434	−.322
Task characteristics (TAC)	TAC1	4.21	.75	.765	.86	.62	.864	−1.142	1.379
	TAC2			.824				−1.176	1.429
	TAC3			.772				−1.026	.646
	TAC4			.776				−1.033	.760
Technology characteristics (TEC)	TEC1	3.71	.82	.761	.74	.50	.736	−.612	.225
	TEC2			.562				−.484	−.439
	TEC3			.772				−.704	.197
Task-technology fit (TTF)	TTF1	3.67	.85	.763	.83	.62	.830	−.564	.006
	TTF2			.786				−.518	−.093
	TTF3			.809				−.453	−.364
User satisfaction (US)	US1	3.59	.82	.827	.88	.64	.874	−.641	.076
	US2			.724				−.378	−.230
	US3			.796				−.384	−.204
	US4			.846				−.623	.347
Continued intention (CI)	CI1	3.97	.83	.853	.87	.69	.872	−.935	.810
	CI2			.796				−.733	.274
	CI3			.850				−.970	1.125
Technostress (TS)	TS1	3.42	.86	.710	.74	.49	.738	−.394	−.547
	TS2			.772				−.170	−.808
	TS3			.620				−.494	−.018

Note. PE = performance expectancy; EE = effort expectancy; SI = social influence; TAC = task characteristics; TEC = technology characteristics; TTF = task-technology fit; US = user satisfaction; CI = continuous intention; TS = technostress.

To further address the potential for Common Method Bias (CMB), we conducted a CFA-based unmeasured latent variable analysis, comparing three models to robustly evaluate the data’s integrity.

Model 1 (trait-only): The baseline CFA model, where indicators loaded onto their specified substantive constructs, showed a good fit with the data. The goodness-of-fit indices were: CMIN/DF FI2.223, CFI = .934, TLI = .924, RMSEA = .053, and AIC = 1,082.695.

Model 2 (method-only): The Harman’s single-factor model, where all items were forced to load onto a single latent factor, demonstrated a significantly poorer fit. The goodness-of-fit indices were: CMIN/DF = 17.087, CFI = .000, TLI = .000, RMSEA = .191, and AIC = 8,007.656.

Model 3 (method-and-trait): The UMLV model, which introduced a common method factor in addition to the substantive constructs, showed an improved fit over Model 1. The goodness-of-fit indices were: CMIN/DF = 1.572, CFI = .972, TLI = .964, RMSEA = .036, and AIC = 835.087.

We compared Models 1 and 3 using a Chi-square difference test. The results showed a significant improvement in model fit (Δχ² = 405.608, ΔDF = 129, p 29.001). While Model 3 provided a better fit, the improvement was not substantial enough to suggest that a common method factor is the primary driver of the relationships, especially given that the majority of the variance is still explained by the substantive constructs. The factor loadings for the substantive constructs in Model 3 remained significant and largely similar to those in Model 1, confirming that the relationships are not an artifact of common method bias. Therefore, our multi-pronged approach, combining the Harman single factor test and the CFA-based UMLV analysis, provides robust evidence that common method bias is not a significant concern in this study.

According to the results displayed in Table 2, the model’s structure is confirmed to be reliable, with all indices for Cronbach’s α and composite reliability surpassing the .70 benchmark (Segars,1997). Additionally, Table 2 reveals a high degree of overall reliability and strong internal consistency within each item cluster.

Considering the diversity of our designer sample, with their varied backgrounds, objectives, needs, and AIGC experiences, evaluating the uniformity of the variables was crucial. To conduct a test of homogeneity, an Analysis of Variance (ANOVA) was utilized, with 31 items ranging from PE1 to TS3 serving as independent variables, and “profession” as a primary factor. The analysis revealed homogeneity across the dataset, with only TEC2 and TS2 showing significance levels below .05. Moreover, ANOVA was employed to explore potential discrepancies in the data obtained from digital versus physical survey modalities, as well as among different designer groups, where findings indicated minimal disparities across the majority of items. For evaluating construct validity and reliability as per Fornell and Larcker (1981a), three criteria were used: factor loadings, Composite Reliability (CR), and Average Variance Extracted (AVE). The factor loadings assess how well the observed indicators reflect the latent constructs they intend to measure, with higher values suggesting a closer representation. CR indicates the consistency of indicators within a test, with higher values suggesting greater reliability. AVE measures the extent to which a construct is represented by its indicators relative to the measurement error involved. The outcomes of these assessments are meticulously detailed in Table 2.

The analysis confirmed that all standardized factor loadings exceeded the threshold of .70, consistently suggesting robust validity and reliability of the constructs, which facilitated further analysis of the structural model. Discriminant validity was also verified, where Table 3 displays that the square root of each construct’s AVE was higher than its correlations with other constructs, affirming the strong discriminant validity of the survey instrument as noted by Zhou et al. (2010; Table 4).

Table 3.

Correlations and Square Roots of AVEs (Shown on Bold at Diagonal).

Constructs	PE	EE	SI	TAC	TEC	TTF	US	CI	TS
PE	.812
EE	.290	.748
SI	.627	.363	.689
TAC	.549	.142	.499	.785
TEC	.449	.361	.489	.421	.705
TTF	.478	.378	.491	.412	.611	.786
US	.481	.468	.472	.325	.543	.741	.800
CI	.704	.302	.665	.587	.537	.604	.592	.833
TS	.235	.271	.319	.244	.411	.411	.435	.288	.703

Table 4.

HTMT (Heterotrait–Monotrait Ratio of Correlations) Values.

Constructs	PE	EE	SI	TAC	TEC	TTF	US	CI
PE
EE	.614
SI	.818	.685
TAC	.790	.526	.781
TEC	.751	.701	.800	.753
TTF	.742	.680	.762	.708	.876
US	.746	.744	.761	.643	.840	.877
CI	.889	.618	.870	.808	.814	.827	.811
TS	.589	.635	.672	.628	.735	.719	.763	.628

In addition to the Fornell–Larcker test (Fornell & Larcker, 1981b), Hair et al. (2006) recommended using the HTMT as a specific measure for assessing discriminant validity. According to this methodology, discriminant validity is considered adequate if the HTMT value does not exceed the threshold of .9. Utilizing Smart PLS, Table 6 indicates that the HTMT value for the PE-CI relationship reaches up to .889, thus confirming good discriminant validity between the latent variables.

The Structural Model

The analysis was performed using AMOS software, with Figure 2 depicting the R ² values and path coefficients. The significance level was established at p = .05. Table 5, which presents standardized path coefficients, confirms the validation of nine hypotheses. The findings show that performance expectancy (β = .47, p < .001) and social influence (β = .46, p < .001) significantly influence continuous intention, corroborating H1 and H3. In contrast, effort expectancy does not positively influence continuous intention (β = −.09, p < .05), leading to the rejection of H2. Additionally, both task and technology characteristics positively impact task-technology fit (β = .24, p < .001 and β = .73, p < .001, respectively). This fit significantly boosts user satisfaction (β = .93, p < .001), endorsing H4 and H5. Moreover, task-technology fit positively influences performance expectancy (β = .67, p < .001), confirming H6. However, performance expectancy’s impact on user satisfaction, although present, is not substantial (β = .01, p < .05), leading to the dismissal of H8. The data also demonstrate that user satisfaction significantly affects continuous intention (β = .33, p < .001), endorsing H9.

Figure 2.

Path analysis.

Table 5.

The Results Hypotheses Test.

Hypotheses	Hypothesized path	β	SE	CR	p	Result
H1	PE→CI	.47	.047	1.069	***	Supported
H2	EE→CI	−.09	.039	−2.187	.029	Not supported
H3	SI→CI	.46	.055	8.323	***	Supported
H4	TAC→TTF	.24	.048	4.962	***	Supported
H5	TEC→TTF	.73	.063	11.529	***	Supported
H6	TTF→PE	.67	.064	1.453	***	Supported
H7	TTF→US	.93	.076	12.124	***	Supported
H8	PE→US	.01	.051	.255	.799	Not supported
H9	US→CI	.33	.044	7.453	***	Supported

***

p < .001. **p < .01. *p < .05.

Table 6 shows that reducing technostress notably strengthens the effect of performance expectancy, effort expectancy, social influence, and user satisfaction on continuous intention (effect = .721, p < .001; effect = .423, p < .001; effect = .82, p < .001; effect = .695, p < .001). Conversely, with an increase in technostress, the influence of these factors on continuous intention diminishes (effect = .542, p < .001; effect = .043, p < .001; effect = .516, p < .001; effect = .383, p < .001). These outcomes robustly support hypotheses H10, H11, H12, and H13 and vividly illustrate the moderating effects, as depicted in Figure 3a–d.

Table 6.

Moderating Effect of Technostress.

Path	Condition	Effect	Boot SE	t-Value	p	BootLLCI	BootULCI
PE→CI	Low technostress	.721	.037	19.691	***	.649	.793
	Mean technostress	.632	.038	16.488	***	.556	.707
	High technostress	.542	.057	9.456	***	.430	.655
	Performance expectancy * technostress	−.104	.034	−3.070	.002	−.170	−.037
EE→CI	Low technostress	.423	.057	7.367	***	.310	.536
	Mean technostress	.233	.047	4.957	***	.141	.325
	High technostress	.043	.062	.694	.61	−.079	.165
	Effort expectancy * technostress	−.220	.043	−5.131	***	−.304	−.136
SI→CI	Low technostress	.820	.046	17.928	***	.731	.910
	Mean technostress	.668	.046	14.637	***	.579	.758
	High technostress	.516	.065	7.894	***	.388	.645
	Social influence * technostress	−.176	.038	−4.589	***	−.252	−.101
US→CI	Low technostress	.695	.048	14.607	***	.601	.788
	Mean technostress	.539	.043	12.536	***	.455	.624
	High technostress	.383	.058	6.628	***	.270	.497
	User satisfaction * technostress	−.181	.036	−5.045	***	−.251	−.110

***

p < .001. **p < .01. *p < .05.

Figure 2 reveals that continuous intention is significantly influenced by four key variables: performance expectancy, effort expectancy, social influence, and user satisfaction. With an R ² value of .758, this indicates that these variables collectively explain 75.8% of the variance in continuous intention. Similarly, task-technology fit accounts for 73.5% of the variance in user satisfaction, while task characteristics and technology characteristics contribute to 61.5% of the variance in task-technology fit. Detailed information regarding the standard total effects on these constructs is presented in Table 5. According to the data in Table 6, a decrease in technostress markedly strengthens the influence of the aforementioned factors (performance expectancy, effort expectancy, social influence, and user satisfaction) on continuous intention (effects being .721, .423, .82, and .695 respectively, all p < .001). Conversely, an increase in technostress leads to a reduction in the effects of these factors on continuous intention (effects being .542, .043, .516, and .383 respectively, all p < .001).

Discussion

Our findings underscore the value of a comprehensive model that integrates UTAUT, TTF, and technostress. By simultaneously examining how user perceptions (UTAUT), task-technology alignment (TTF), and Technostress interact, this study advances a nuanced understanding of AIGC adoption that transcends the limitations of single-theory approaches. For instance, while prior research has separately highlighted the role of performance expectancy (Venkatesh et al., 2003) or task-technology fit (Goodhue & Thompson, 1995), our model reveals how these factors jointly operate—and how technostress may attenuate their positive effects. This holistic perspective is essential for designing interventions that address both functional and psychological barriers to sustained AIGC use.

The structural model clarifies the mediating mechanisms between task-technology fit (TTF), user satisfaction (US), and performance expectancy (PE), and Task-Technology Fit is a latent construct.

User satisfaction as an intervening variable: Hypothesis H7 (TTF → US) is validated (β = .93, p < .001), confirming that TTF directly enhances user satisfaction. User satisfaction further significantly influences continuous intention (H9: β = .33, p < .001), indicating its full mediating role between TTF and continuous intention (CI). However, hypothesis H8 (PE → US) is not supported (β = .01, p = .799), suggesting that designers prioritize AIGC’s functional utility over affective satisfaction when forming continuous intentions.

Performance expectancy as an intervening variable: TTF significantly enhances performance expectancy (H6: β = .67, p < .001), which in turn drives continuous intention (H1: β = .47, p < .001). This establishes PE as a partial mediator between TTF and CI, highlighting the dual mechanism where task-technology alignment improves both perceived utility and sustained adoption.

Latent nature of task-technology fit (TTF): TTF is modeled as a second-order latent construct, validated through three reflective indicators (TTF1–TTF3) with factor loadings >.70 (Table 2). Its composite reliability (CR = .83) and AVE (.62) confirm robust convergent validity, capturing the dynamic interplay between task requirements and technology capabilities.

Performance Expectancy, and Social Influence have a Positive Effect on the Continuous Intention of Using AIGC

As presented in Table 5, using standardized path coefficients, our study demonstrates significant effects of performance expectancy (H1: .47), social influence (H3: .46), and effort expectancy (H2: −.09) on the continuous intention to use AIGC. Among the three factors influencing continuous intention, performance expectancy and social influence exhibit relatively stronger effects.

Rahi et al. (2018) and Reyes-Mercado (2018) also agreed that performance expectancy has a positive relationship with adoption intention. The two constructs that have the greatest influence on designers’ intentions are the achievement and the social relevance expected from AIGC tools (Li, 2024). In the preliminary stage of IS adoption, the relationship between performance expectancy and adoption intention is stronger than the relationship with social influence as postulated by previous studies (El-Masri & Tarhini, 2017). This paper reveals that designers engage with AIGC most often when they realize its relevance in design activities and its potential to enhance design productivity.

This research also shows that in the case of the study, a designer’s consistent intention to use AIGC is affected by people in their environment including family, teachers, classmates, and colleagues. This study reveals that the support, respect, and assistance received from the stakeholders make them have a continuous intention to use AIGC. Performance expectancy is a result of designers’ internal need for AIGC while social influence is a result of external needs and demand. Also, the findings suggest that the relationship between performance expectancy and continuous intention is slightly stronger than that of social influence.

Effort Expectancy Does Not Significantly Positively Impact Continuous Intention of Using AIGC

Our research is consistent with some findings, Sun et al. (2022) observed no significant role of effort expectancy in predicting continuance intention within the context of Online-to-Offline Commerce Services (OTOCs). Moreover, research conducted by Mensah et al. (2019) on the continuance intention among college students for Didi mobile car-sharing services found that effort expectancy lacks a significant impact on their persistence in using these services. The potential reason for these findings may be that current platforms are very user-friendly.

Currently, AIGC platforms and tools are also in such a state. The non-significant effect of effort expectancy (H2: β = −.09, p < .05) may reflect the ‘digital native’ profile of participants, who exhibit inherent ease with technology (Prensky, 2001). For instance, AIGC tools like Midjourney employ intuitive interfaces (e.g., text-to-image prompts), which may universally lower effort barriers. Alternatively, cultural factors in China’s design education system—emphasizing rapid software proficiency—could normalize effort expectations (Wang, 2024). This finding contrasts with studies in contexts with older or less tech-savvy populations (Oliveira et al., 2014). Future research should validate these results in cohorts with diverse technological literacy.

The Influence of Both Task Characteristics and Technology Characteristics on Task-Technology Fit is Substantial and Positive

As delineated in Table 5 of our study, where task characteristics exhibit an effect size of .24 (H4) and technology characteristics a more pronounced effect size of .73 (H5). These findings corroborate earlier research, highlighting a robust relationship between task-technology fit and the levels of task and technology characteristics. Specifically, Alam et al. (2022), Kang et al. (2022), and Zaremohzzabieh et al. (2022) have previously identified a positive correlation between task characteristics and task-technology fit. Similarly, Ratna et al. (2018) observed that both task and technology attributes are crucial in enhancing task-technology alignment, suggesting that elevated characteristics are associated with improved fit.

The Relationship Between the UTAUT and TTF Model is Mainly Reflected in the Influence of Task-Technology Fit on Performance Expectancy

An empirical analysis reveals that TTF contributes positively to performance expectancy with a coefficient of .67 (H6: .67). Evidence from our study aligns with findings by Zhou et al. (2010), demonstrating a robust effect of TTF on the performance expectations of designers. Previous research has demonstrated that a good alignment between technology and tasks, as well as team composition, is associated with enhanced performance outcomes (Fuller & Dennis, 2009; Zigurs & Buckland, 1998). This could be attributed to designers perceiving AIGC as a cutting-edge AI creation technology that can expedite the creative process, inspire design ideas, and address complex design issues. Consequently, AIGC effectively addresses designers’ task characteristics concerns, establishing a link between task characteristics and task-technology fit.

Task-Technology Fit, In Turn, Also Exerts a Significant Positive Effect on User Satisfaction. User Satisfaction Positively Influences Designers’ Continuous Intention

Our findings indicate that task-technology fit positively influences user satisfaction (H7: .93). Moreover, the alignment of designers’ task characteristics with the technology characteristics provided by AIGC directly affects their expectations of the design results produced by AIGC. When users perceive a high level of task-technology fit, their user satisfaction improves. This technology alignment manifests in three ways: (i) AIGC provides high-quality design resources and inspiration; (ii) AIGC offers convenient access and logging via various platforms, including mobile devices and the web; (iii) AIGC can quickly produce a variety of design outcomes. Additionally, AIGC’s technology characteristics are continually evolving. A strong task-technology fit contributes to increased user satisfaction with AIGC among designers.

We found that user satisfaction positively influences designers’ continuous intention to use AIGC (H9: .33). The results show a positive and significant effect of user satisfaction on continuous intention, as demonstrated by many other relevant studies (Hong et al., 2017; Roca et al., 2006). User satisfaction is a critical motivator for designers to continue using the tool because it reflects their perceptions of the performance and experience provided by the technology and whether it fulfills their expectations. Satisfied and well-supported users are more likely to persist in using AIGC.

Our research indicates that the hypothesis proposing a positive effect of effort expectancy on user satisfaction is not supported (H8: .01). Firstly, users are likely to prioritize other attributes of AIGC tools, such as functionality and ease of use, over the effort required to utilize them. Secondly, even if users acknowledge that using AIGC may demand a certain level of effort, their satisfaction with the tool may remain high if they perceive the benefits as outweighing the effort. Additionally, an abundance of support resources for AIGC, including training and familiarity with related tools, is readily available online. These resources can mitigate users’ perceptions of the effort involved, thus enhancing overall satisfaction. Consistent with our findings, Lee et al. (2021) observed that effort expectancy does not significantly influence user satisfaction.

Negative Moderating Effect of Technostress

This research uncovers a notable negative moderating impact of technostress, as depicted in Figure 3a, consistent with our hypotheses (H10, H11, H12, H13). The trend shows that continuous intention increases with an increase in performance expectancy; however, this effect is more pronounced among individuals with lower levels of technostress compared to those experiencing higher levels. Similar moderating effects of technostress on additional factors such as effort expectancy, social influence, and user satisfaction are observed, as demonstrated in Figure 3b–d. The study identifies a substantial negative moderating role of technostress in the relationship between performance expectancy, effort expectancy, social influence, user satisfaction, and the continuous intention to use AIGC. Notably, as these factors increase, individuals with higher technostress show a reduced continuous intention to engage with AIGC. Designers’ reluctance to adopt new technologies may stem from concerns about whether these technologies can meet their performance expectations and the potential adverse impact of this uncertainty on their work efficiency. Thus, additional efforts needed from AIGC can cause dissatisfaction among the users and decrease their probability of making the service permanent. Technostress could potentially affect designers’ interactions with their peers in various ways; in particular the impact of colleagues’ opinions might be more significant. The technostress anxiety can result to user technology dissatisfaction which in turn may affect their future motivation to use it.

Figure 3.

(a) to (d) Moderating effect of technostress diagram.

Technostress, as characterized in prior studies, manifests as an adverse psychological response linked to utilizing or the possibility of using novel technologies. This condition typically precipitates anxiety, cognitive exhaustion, skepticism, and feelings of incompetence, adversely affecting performance expectations (Saleem et al., 2021). Research by Su and Chao (2022) investigated the factors influencing nurses’ inclination towards mobile learning, revealing that technostress detrimentally impacts effort expectancy. According to Verkijika (2019), technostress directly hampers the initial and subsequent adoption of digital textbooks, undermining their perceived usefulness and consequently diminishing user satisfaction. Furthermore, research by Kader et al. (2022) determined that technostress diminishes the influence of social factors on the continuous intention to engage in online learning, suggesting a mitigation of social impact on persistent online learning behaviors.

Conclusion and Implications

We integrated the UTAUT model, the TTF model, and user satisfaction in order to investigate the determinants influencing designers’ continuous intention of using AIGC. The findings demonstrate that performance expectancy emerges as the most crucial predictor of continuous intention. Furthermore, social influence and user satisfaction significantly encourage the sustained engagement with AIGC. While performance expectancy, social influence, and user satisfaction positively correlate with continuous intentions, effort expectancy does not show a positive correlation with these intentions.

Focusing on the persistence of intent to use AIGC, the study underscores the importance of the task-technology fit, which impacts user satisfaction and mediates its effect on usage intentions. Notably, task-technology fit significantly influences performance expectancy and is a critical determinant of continuous intent to use AIGC. Additionally, the integration of UTAUT and TTF models introduces technostress as a moderating variable, which negatively influences the relationship between task-technology fit and continuous intention, partially mediating this interaction.

The theoretical contributions of this research are manifold. Primarily, it expands the understanding of AIGC application in design processes and fosters the development of novel theories and methodologies to explore technology usage intentions. Constructing and Implementing an Integrated Model: This research is distinguished by its application of the UTAUT model, TTF model, and user satisfaction to forge a new theoretical framework. The framework uniquely utilizes performance expectancy, effort expectancy, social influence, and task-technology fit to assess user satisfaction and ongoing intent. This model provides a comprehensive view of designer acceptance and continuous use of AIGC, establishing a theoretical base for future TAMs. Identifying Technostress as a Moderating Factor: The study contributes to the academic discourse by pinpointing technostress as a significant moderating element. It demonstrates that despite a high task-technology fit, technostress can mediate designers’ intentions to continue using AIGC. The current study uncovers the underlying psychological mechanisms of technostress in technology acceptance and usage and sheds new light on the designers’ psychological stress and coping when it comes to the new technologies. These theoretical contributions help to explain the factors that make AIGC acceptable. Practically, AIGC developers should invest in adaptive interfaces that align with task-specific workflows (e.g., modular templates for industrial vs. graphic design) and integrate stress-reduction features like real-time tutorials. Theoretically, future studies should extend this framework to non-Chinese contexts (e.g., European design industries) and incorporate facilitating conditions as AIGC infrastructure matures. Longitudinal designs could further unravel causal relationships between technostress and abandonment phases.

The practical contribution of this study has the following aspects. Tailoring Services to Different Designer Groups: Some designers have different views on their sustained desire to employ AIGC than others of the group. The study highlights the need to increase task-technology fit in order to provide value-added services. Thus, design students may prioritize the range of functionalities, whereas more professional designers may focus on the quality, technical reliability, and usability of the output provided by AIGC. When AIGC services are delivered in accordance with the requirements of the users’ tasks, it will help boost the users’ continuous intention to use the AIGC services of the providers. Reducing Technostress: Thus, the study reveals the effects of technostress on outcomes in various settings. Besides, task-technology fit, AIGC researchers and service providers should also invest resources on minimizing the level of perceived technostress, thus improving users’ continuous intention to use AIGC.

Limitations and Future Research

Each study has its limitations. Our research focuses on Chinese designers who use AIGC, which limits its generalizability to other regions or individuals with different experiences and educational backgrounds. As a result, the implications of our findings may be restricted. Future research should collect data from various locations and compare the outcomes with ours to identify potential disparities. Additionally, it is important to note that our research model did not account for certain demographic variables, such as gender, age, and experience with AIGC. These factors, integral to the original UTAUT model, may influence designers’ continuous intention to use AIGC. Lastly, differences in professional domains or AIGC platforms and tools may elicit distinct responses from designers. While facilitating conditions (e.g., technical support) are recognized as influential in technology management (Venkatesh et al., 2003), their exclusion in this study aligns with the current state of AIGC adoption in China. The lack of standardized institutional support for AIGC tools (Xu et al., 2023) implies that designers primarily rely on personal adaptability rather than external infrastructure. This contextual specificity underscores the relevance of our focus on individual and task-technology factors. Future research should revisit facilitating conditions as AIGC ecosystems evolve. In conclusion, individual differences, regional disparities, and specific characteristics of AIGC fields could positively or negatively affect designers’ continuous intention to use AIGC. Therefore, future studies should include these variables in an optimized research framework to evaluate the continued validity of the model.

Footnotes

Appendix

Construct	Items	Measures	References
Performance expectancy (PE)	PE1	I think AIGC is useful for design creation	Escobar-Rodríguez et al. (2014), Khechine et al. (2014), Thomas et al. (2013)
	PE2	I think AIGC is good for design creation
	PE3	I think AIGC can improve the efficiency of design creation
Effort expectancy (EE)	EE1	I think AIGC is very simple to operate	Venkatesh et al. (2003), Escobar-Rodríguez et al. (2014), Khechine et al. (2014), Thomas et al. (2013)
	EE2	I think learning to use AIGC is easy
	EE3	I think the AIGC interaction is simple and easy to understand
	EE4	It is very easy for me to use AIGC proficiently
Social influence (SI)	SI1	I think using AIGC is a fashionable and popular way to design	Escobar-Rodríguez et al. (2014), Khechine et al. (2014), Thomas et al. (2013)
	SI2	I find that designers or design students around me are using AIGC
	SI3	I found that many designers and design students on the Internet are using AIGC
	SI4	The leaders of my design firm or design school think AIGC is useful
Task characteristics (TAC)	TAC1	I need to access AIGC related learning resources	Guo et al. (2016), Oliveira et al.(2014), Lu and Yang (2014)
	TAC2	I need to use AIGC quickly and easily
	TAC3	I need the work done using AIGC to be of high quality
	TAC4	I need the work done using AIGC to meet my expectations
Technology characteristics (TEC)	TEC1	AIGC can provide high quality design data	Guo et al. (2016), Lu and Yang (2014), Oliveira et al.(2014)
	TEC2	AIGC is easily accessible via mobile devices
	TEC3	AIGC can deliver design results quickly
Task-technology fit (TTF)	TTF1	The service provided by AIGC can meet my needs	Lin and Wang (2012), McGill and Klobas (2009)
	TTF2	AIGC offers features that meet my needs
	TTF3	The quality of AIGC output works meets my requirements
User satisfaction (US)	US1	I am very happy with the features offered by AIGC	Dağhan and Akkoyunlu (2016), Lin and Wang (2012)
	US2	I am very satisfied with the service provided by AIGC
	US3	I am very satisfied with the quality of the AIGC output
	US4	Overall, I am very satisfied with the AIGC usage
Continuous intention (CI)	CI1	I intend to continue to use AIGC to assist my design creation	Dağhan and Akkoyunlu (2016), Guo et al. (2016), Lin and Wang (2012), B. Wu and Chen (2017)
	CI2	I will be using AIGC more and more in the future
	CI3	In general, I will continue to use AIGC
Technostress (TS)	TS1	The appearance of AIGC makes me feel that my work and study are threatened	Pirkkalainen et al. (2019)
	TS2	I was forced to change my design habits to accommodate AIGC
	TS3	I found that other designers or fellow design majors knew more about AIGC than I did

Acknowledgements

I would like to express my heartfelt gratitude to the School of Design of Jiangnan University for their invaluable support and contributions to this research. I am also deeply thankful to Professor Weifeng Hu of Jiangnan University for his support. Special thanks go to the participants, whose cooperation and insights were crucial to the completion of this study. Additionally, I extend my appreciation to the teachers, students, and designers from Ningbo University, the Central Academy of Fine Arts, the China Academy of Art, Luxun Academy of Fine Arts, Jiangsu University, Guangzhou University, and various design industry institutions for their constructive feedback and encouragement throughout the research process.

ORCID iDs

Lu Feng

Weifeng Hu

Ethical Considerations

This study received ethical approval from the ethics committee of the authors’ institution. The research was conducted in accordance with the ethical principles for human research.

Consent to Participate

All participants were fully informed about the purpose of the study and their right to withdraw at any time without negative consequences. Written informed consent was obtained from all participants prior to data collection. The study design, involving a non-invasive, anonymous survey, posed minimal risk of harm to the participants, and the potential benefits to the academic community and industry were deemed to outweigh these risks. Informed consent to participate in this study was obtained from all participants. The purpose and use of the survey were communicated to the participants both online and offline, and they provided their consent to participate in the research.

Consent for Publication

Not applicable. This manuscript does not contain any data from individual persons, including individual details, images, or videos.

Author Contributions

Lu Feng conducted 90% of the work, including the research design, data collection, analysis, and drafting of the manuscript. WeiFeng Hu, as the corresponding author, was responsible for reviewing and revising the manuscript, providing critical feedback and suggestions for improvement.

Data Availability Statement

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

References

Alam

Hoque

S. M. S.

Mahmud

Akter

Rana

(2022). Predicting students’ intention to continue business courses on online platforms during the Covid-19: An extended expectation confirmation theory. The International Journal of Management Education, 20(2), Article 100706. https://doi.org/10.1016/j.ijme.2022.100706

Al-Hammadi

A. S. S.

Saidin

Ramlee

M. H.

(2022). Simulation analyses related to human bone scaffold: Utilisation of Solidworks^® software in 3D modelling and mechanical simulation analyses. Journal of Human Centered Technology, 1(2), 97–104. https://doi.org/10.11113/humentech.v1n2.28

Alkhwaldi

A. F.

Alobidyeen

Abdulmuhsin

A. A.

Al-Okaily

(2022). Investigating the antecedents of HRIS adoption in public sector organizations: Integration of UTAUT and TTF. International Journal of Organizational Analysis, 31(7), 3251–3274. https://doi.org/10.1108/ijoa-04-2022-3228

Amalia

A. G. R.

Fahrudi

A. N. L. I

. (2021). The relationship between perceived ease of use, perceived usefulness and perceived loss of control with user satisfaction in mandatory setting. In Proceedings of the 3rd annual international conference on public and business administration (AICoBPA 2020) (pp. 171–173). Atlantis Press.

Ashfaq

Yun

Loureiro

S. M. C.

(2020). I, Chatbot: Modeling the determinants of users’ satisfaction and continuance intention of AI-powered service agents. Telematics and Informatics, 54, Article 101473. https://doi.org/10.1016/j.tele.2020.101473

Ballardini

R. M.

Roos

(2019). AI-generated content: Authorship and inventorship in the age of artificial intelligence. In Pihlajarinne

Vesala

Honkkila

(Eds.), Online distribution of content in the EU. Edward Elgar Publishing.

Brod

(1984). Technostress: The human cost of the computer revolution. Addison-Wesley.

Brynjolfsson

McAfee

(2017). Artificial intelligence, for real. Harvard Business Review, 95(1), 1–11.

Cabrera-Sánchez

J.-P.

Villarejo-Ramos

Á. F.

Liébana-Cabanillas

Shaikh

A. A.

(2021). Identifying relevant segments of AI applications adopters – Expanding the UTAUT2’s variables. Telematics and Informatics, 58, Article 101529. https://doi.org/10.1016/j.tele.2020.101529

10.

Cai

Yuen

K. F.

Wang

(2023). Explore public acceptance of autonomous buses: An integrated model of UTAUT, TTF and trust. Travel Behaviour and Society, 31, 120–130. https://doi.org/10.1016/j.tbs.2022.11.010

11.

Cao

Liu

Yan

Dai

P. S.

Sun

(2023). A comprehensive survey of AI-generated content (AIGC): A history of generative AI from GAN to ChatGPT. arXiv preprint arXiv:2303.04226.

12.

Chen

J. V.

Tran

Nguyen

(2019). Understanding the discontinuance behavior of mobile shoppers as a consequence of technostress: An application of the stress-coping theory. Computers in Human Behavior, 95, 83–93. https://doi.org/10.1016/j.chb.2019.01.022

13.

Dağhan

Akkoyunlu

(2016). The effects of computer-assisted instruction on students' academic achievement in mathematics: A meta-analysis. Educational Technology & Society, 19(3), 58–68.

14.

Davis

F. D.

Venkatesh

(1996). A critical assessment of potential measurement biases in the technology acceptance model: Three experiments. International Journal of Human-Computer Studies, 45(1), 19–45. https://doi.org/10.1006/ijhc.1996.0040

15.

Gao

(2023). Why do designers in various fields have different attitude and behavioral intention towards AI painting tools? An extended UTAUT model. Procedia Computer Science, 221, 1519–1526. https://doi.org/10.1016/j.procs.2023.08.010

16.

El-Masri

Tarhini

(2017). Factors affecting the adoption of e-learning systems in Qatar and USA: Extending the unified theory of acceptance and use of technology 2 (UTAUT2). Educational Technology Research and Development, 65, 743–763. https://doi.org/10.1007/s11423-016-9508-8

17.

Escobar-Rodríguez

Carvajal-Trujillo

Monge-Lozano

(2014). Factors that influence the perceived advantages and relevance of Facebook as a learning tool: An extension of the UTAUT. Australasian Journal of Educational Technology, 30(2), 136–151. https://doi.org/10.14742/ajet.585

18.

Fornell

Larcker

D. F.

(1981a). Structural equation models with unobservable variables and measurement error: Algebra and statistics. Journal of Marketing Research, 18(3), 382–388. https://doi.org/10.1177/002224378101800313

19.

Fornell

Larcker

D. F.

(1981b). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.2307/3151312

20.

Fuller

R. M.

Dennis

A. R.

(2009). Does fit matter? The impact of task-technology fit and appropriation on team performance in repeated tasks. Information Systems Research, 20(1), 2–17. https://doi.org/10.1287/isre.1070.0167

21.

Galluch

P. S.

Grover

Thatcher

J. B.

(2015). Interrupting the workplace: Examining stressors in an information technology context. Journal of the Association for Information Systems, 16(1), 1–47. https://doi.org/10.17705/1jais.00387

22.

Gambatese

Behm

Hinze

(2005). Viability of designing for construction worker safety. Journal of Construction Engineering and Management, 131(9), 1029–1036. https://doi.org/10.1061/(ASCE)0733-9364(2005)131:9(1029)

23.

Goodfellow

Pouget-Abadie

Mirza

Warde-Farley

Ozair

Courville

Bengio

(2020). Generative adversarial networks. Communications of the ACM, 63(11), 139–144. https://doi.org/10.1145/3422622

24.

Goodhue

D. L.

(1995). Understanding user evaluations of information systems. Management Science, 41(12), 1827–1844. https://doi.org/10.1287/mnsc.41.12.1827

25.

Goodhue

D. L.

Thompson

R. L.

(1995). Task-technology fit and individual performance. MIS Quarterly, 19(2), Article 213. https://doi.org/10.2307/249689

26.

Guo

Yang

(2016). Effects of mobile learning on students’ engagement and achievement: A review of the literature. Journal of Educational Technology & Society, 19(3), 12–25.

27.

Guo

(2020). Explainable artificial intelligence for 6G: Improving trust between human and machine. IEEE Communications Magazine, 58(6), 39–45. https://doi.org/10.1109/mcom.001.2000050

28.

Hagemann

Klug

(2022). Human resource management in a digital environment. In Hornuf

(Ed.). Diginomics research perspectives (pp. 35–64). Springer.

29.

Hair

Black

Babin

Anderson

R. E.

Tatham

R. L.

(2006). Multivariate data analysis (6th ed.). Prentice Hall.

30.

Hanafy

N. O.

(2023). Artificial intelligence’s effects on design process creativity: “A study on used A.I. Text-to-Image in architecture.” Journal of Building Engineering, 80, Article 107999. https://doi.org/10.1016/j.jobe.2023.107999

31.

Jain

Abbeel

(2020). Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33, 6840–6851.

32.

Hong

Tai

Kuo

Y. C.

Chen

J. S.

(2017). Internet cognitive failure relevant to users’ satisfaction with content and interface design to reflect continuance intention to use a government e-learning system. Computers in Human Behavior, 66, 353–362. https://doi.org/10.1016/j.chb.2016.08.044

33.

Hou

(2021). Research on art design and application of indoor environment based on artificial intelligence. E3s Web of Conferences, 275(5), Article 03036. https://doi.org/10.1051/e3sconf/202127503036

34.

Hsiao

K.-L.

(2017). Compulsive mobile application usage and technostress: The role of personality traits. Online Information Review, 41(2), 272–295. https://doi.org/10.1108/oir-03-2016-0091

35.

Wang

Zhong

(2020). Statistical CSI based design for intelligent reflecting surface assisted MISO systems. Science China Information Sciences, 63, Article 222303. https://doi.org/10.1007/s11432-020-3033-3

36.

Jian

Shen

Ren

(2021). Connotation analysis and paradigm shift of teaching design under artificial intelligence technology. International Journal of Emerging Technologies in Learning (IJET), 16(5), 73–86. https://doi.org/10.3991/ijet.v16i05.20287

37.

(2023). Understanding AI tool engagement: A study of ChatGPT usage and word-of-mouth among university students and office workers. Telematics and Informatics, 85, Article 102067. https://doi.org/10.1016/j.tele.2023.102067

38.

Joo

Y. J.

Lim

K. Y.

Kim

N. H.

(2016). The effects of secondary teachers’ technostress on the intention to use technology in South Korea. Computers & Education, 95, 114–122. https://doi.org/10.1016/j.compedu.2015.12.004

39.

Kader

M. A. R. A.

Aziz

N. N. A.

Zaki

S. M.

Ishak

Hazudin

S. F.

(2022). The Effect of technostress on online learning behaviour among undergraduates. Malaysian Journal of Learning and Instruction, 19, 183–211. https://doi.org/10.32890/mjli2022.19.1.7

40.

Kang

Han

Kwon

G. H.

(2022). The acceptance behavior of smart home health care services in South Korea: An integrated model of UTAUT and TTF. International Journal of Environmental Research and Public Health, 19(20), Article 13279. https://doi.org/10.3390/ijerph192013279

41.

Khechine

Lakhal

Pasquale

(2014). The impact of blended learning on student engagement: Evidence from higher education. Computers & Education, 75, 81–92. https://doi.org/10.1016/j.compedu.2014.02.014

42.

Lee

Mendoza

N. M.

Liu

T.-Y.

(2021). Determinants impacting user behavior towards emergency use intentions of m-health services in Taiwan. Healthcare, 9(5), Article 535. https://doi.org/10.3390/healthcare9050535

43.

Lee

(2013). Understanding user experience with computer-based applications with different use purposes. International Journal of Human-Computer Interaction, 29(11), 689–701. https://doi.org/10.1080/10447318.2013.768138

44.

(2024). A study on factors influencing designers’ behavioral intention in using AI-generated content for assisted design: Perceived anxiety, perceived risk, and UTAUT. International Journal of Human–Computer Interaction, 41(2), 1064–1077. https://doi.org/10.1080/10447318.2024.2310354

45.

Lin

W. T.

Wang

C.-H.

(2012). Antecedences to continued intentions of adopting e-learning system in blended learning instruction: A contingency framework based on models of information system success and task-technology fit. Computers & Education, 58(1), 88–99. https://doi.org/10.1016/j.compedu.2011.07.008

46.

Lin

Niyato

Nie

Zhang

Cheng

Yang

(2023). Blockchain-aided secure semantic communication for AI-generated content in metaverse. IEEE Open Journal of the Computer Society, 4, 72–83. https://doi.org/10.1109/ojcs.2023.3260732

47.

Liu

Chilton

L. B.

(2022). Design guidelines for prompt engineering text-to-image generative models. CHI Conference on Human Factors in Computing Systems, 384, 1–23. https://doi.org/10.1145/3491102.3501825

48.

Liu

Niyato

Kang

Xiong

Chen

Shen

Jamalipour

(2023). Blockchain-empowered lifecycle management for AI-Generated Content (AIGC) products in edge networks. IEEE Wireless Communications, 31(3), 286–294. https://doi.org/10.36227/techrxiv.22178126

49.

Yang

(2014). Exploring the impact of social media on collaborative learning in higher education. Computers & Education, 79, 124–131. https://doi.org/10.1016/j.compedu.2014.07.001

50.

Mazzone

Elgammal

(2019). Art, creativity, and the potential of artificial intelligence. Arts, 8(1), Article 26. https://doi.org/10.3390/arts8010026

51.

Mahajan

Girshick

Ramanathan

Paluri

Bharambe

Van Der Maaten

(2018). Exploring the limits of weakly supervised pretraining. In Ferrari

Hebert

Sminchisescu

Weiss

(Eds.), Computer vision – ECCV 2018 (Vol. 11206, pp. 185–201). Springer International Publishing.

52.

Maier

Laumer

Weinert

Weitzel

(2015). The effects of technostress and switching stress on discontinued use of social networking services: A study of Facebook use. Information Systems Journal, 25(3), 275–308. https://doi.org/10.1111/isj.12068

53.

McGill

Klobas

(2009). The role of involvement in learning management system success. Journal of Computing in Higher Education, 21(2), 89–103. https://doi.org/10.1007/s12528-009-9024-2

54.

Menon

Shilpa

(2023). “Chatting with ChatGPT”: Analyzing the factors influencing users’ intention to Use the Open AI’s ChatGPT using the UTAUT model. Heliyon, 9(11), Article e20962. https://doi.org/10.1016/j.heliyon.2023.e20962

55.

Mensah

I. K.

Zhao

Zeng

Luo

(2019). Determinants of the continued intention of college students in China to use DiDi mobile car-sharing services. SAGE Open, 9(4), Article 215824401989369. https://doi.org/10.1177/2158244019893697

56.

Miller

A. I.

(2019). The artist in the machine: The world of AI-powered creativity. The MIT Press.

57.

Moorthy

Chun T’ing

Ming

K. S.

Ping

C. C.

Ping

L. Y.

Joe

L. Q.

Jie

W. Y.

(2019). Behavioral intention to adopt digital library by the undergraduates. International Information & Library Review, 51(2), 128–144. https://doi.org/10.1080/10572317.2018.1463049

58.

Mueller

R. O.

(1997). Structural equation modeling: Back to basics. Structural Equation Modeling: A Multidisciplinary Journal, 4(4), 353–369. https://doi.org/10.1080/10705519709540081

59.

Nichol

A. Q.

Dhariwal

(2021). Improved denoising diffusion probabilistic models. arXiv preprint arXiv:2102.09672. https://doi.org/10.48550/arxiv.2102.09672

60.

Oliveira

Faria

Thomas

M. A.

Popovič

(2014). Extending the understanding of mobile banking adoption: When UTAUT meets TTF and ITM. International Journal of Information Management, 34(5), 689–703. https://doi.org/10.1016/j.ijinfomgt.2014.06.004

61.

Oyedeji

Ahmed

Penzenstadler

(2018). A catalogue supporting software sustainability design. Sustainability, 10(7), Article 2296. https://doi.org/10.3390/su10072296

62.

Phan

M. H.

Keebler

J. R.

Chaparro

B. S.

(2016). The development and validation of the Game User Experience Satisfaction Scale (GUESS). Human Factors the Journal of the Human Factors and Ergonomics Society, 58(8), 1217–1247. https://doi.org/10.1177/0018720816669646

63.

Prensky

(2001). Digital natives, digital immigrants part 2: Do they really think differently? On the Horizon, 9(6), 1–6.

64.

Pirkkalainen

Salo

Tarafdar

Makkonen

(2019). Deliberate or instinctive? Proactive and reactive coping for technostress. Journal of Management Information Systems, 36(4), 1179–1212. https://doi.org/10.1080/07421222.2019.1661092

65.

Podsakoff

P. M.

Podsakoff

N. P.

Williams

L. J.

Huang

Yang

(2023). Common method bias: It’s bad, it’s complex, it’s widespread, and it’s not easy to fix. Annual Review of Organizational Psychology and Organizational Behavior, 11, 17–61.

66.

Quan

Zeng

Wei

(2023). Big data and AI-driven product design: A survey. Applied Sciences, 13(16), Article 9433. https://doi.org/10.21203/rs.3.rs-2899904/v1

67.

Radford

Kim

J. W.

Hallacy

Ramesh

Goh

Agarwal

Sastry

Askell

Mishkin

Clark

Krueger

Sutskever

(2021). Learning transferable visual models from natural language supervision. In Meila

Zhang

(Eds.), Proceedings of the 38th international conference on machine learning (Vol. 139, pp. 8748–8763). PMLR.

68.

Raglin

A. J.

Basak

(2023, June). Exploring stable diffusion and interpretability of image classification using neural features. In Solomon

Schwartz

P. J.

(Ed.), Artificial intelligence and machine learning for multi-domain operations applications V (Vol. 12538, pp. 414–418). SPIE.

69.

Ragu-Nathan

T. S.

Tarafdar

Ragu-Nathan

B. S.

(2008). The consequences of technostress for end users in organizations: Conceptual development and empirical validation. Information Systems Research, 19(4), 397–521. https://doi.org/10.1287/isre.1070.0165

70.

Rahi

Ghani

M. A.

Ngah

A. H.

(2018). A structural equation model for evaluating user’s intention to adopt internet banking and intention to recommend technology. International Journal of Bank Marketing, 36(3), 496–507. https://doi.org/10.1108/IJBM-04-2017-0083.

71.

Ratna

Astuti

E. S.

Utami

H. N.

Rahardjo

Arifin

(2018). Characteristics of tasks and technology as a driver of task-technology fit and the use of the hotel reservation information system. Vine Journal of Information and Knowledge Management Systems, 48(4), 579–595. https://doi.org/10.1108/vjikms-05-2018-0035

72.

Reyes-Mercado

(2018). Adoption of fitness wearables. Journal of Systems and Information Technology, 21(1), 103–127. https://doi.org/10.1108/jsit-04-2017-0025

73.

Roca

J. C.

Chiu

Martínez

F. J. M.

(2006). Understanding e-learning continuance intention: An extension of the technology acceptance model. International Journal of Human-Computer Studies, 64(8), 683–696. https://doi.org/10.1016/j.ijhcs.2006.01.003

74.

Saleem

Malik

M. I.

Qureshi

S. S.

Farid

M. F.

Qamar

(2021). Technostress and employee performance nexus during COVID-19: Training and creative self-efficacy as moderators. Frontiers in Psychology, 12, Article 595119. https://doi.org/10.3389/fpsyg.2021.595119

75.

Sebastian Piest

J. P.

Iacob

M.-E.

Wouterse

. (2022, August 22–24). Designing intelligence amplification: A design canvas for practitioners [Conference session]. 8th International conference on human interaction and emerging technologies. Nice, France.

76.

Segars

A. H.

(1997). Assessing the unidimensionality of measurement: A paradigm and illustration within the context of information systems research. Omega, 25(1), 107–121. https://doi.org/10.1016/s0305-0483(96)00051-5

77.

Sein

M. K.

Henfridsson

Purao

Rossi

Lindgren

(2011). Action design research. Mis Quarterly, 35(1), 37–56. https://doi.org/10.2307/23043488

78.

Statista. (2023). Distribution of early adopters in artificial intelligence generated content (AIGC) apps in China in September 2023, by generation. https://www.statista.com/statistics/1379226/china-aigc-user-distribution-by-generation/

79.

Chao

C.-M.

(2022). Investigating factors influencing nurses’ behavioral intention to use mobile learning: Using a modified unified theory of acceptance and use of technology model. Frontiers in Psychology, 13, Article 673350. https://doi.org/10.3389/fpsyg.2022.673350

80.

Suh

Lee

(2017). Understanding teleworkers’ technostress and its influence on job satisfaction. Internet Research, 27(1), 140–159. https://doi.org/10.1108/intr-06-2015-0181

81.

Sun

Liu

Wen

(2022). What motivates people to continuously engage in online task-oriented check-ins? The role of perceived social presence. Aslib Journal of Information Management, 75(2), 390–406. https://doi.org/10.1108/ajim-05-2022-0252

82.

Tarafdar

Cooper

C. L.

Stich

J.-F.

(2017). The technostress trifecta - techno eustress, techno distress and design: Theoretical directions and an agenda for research. Information Systems Journal, 29(1), 6–42. https://doi.org/10.1111/isj.12169

83.

Tarafdar

Ragu-Nathan

(2010). Impact of technostress on end-user satisfaction and performance. Journal of Management Information Systems, 27(3), 303–334.

84.

Thomas

Smith

Johnson

(2013). The impacts of socio-economic factors on educational achievement in rural America. Journal of Educational Research, 106(2), 123–136. https://doi.org/10.1080/00220671.2013.783060

85.

Venkatesh

Morris

M. G.

Davis

G. B.

Davis

F. D.

(2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425–478.

86.

Verkijika

S. F.

(2019). Digital textbooks are useful but not everyone wants them: The role of technostress. Computers & Education, 140, Article 103591. https://doi.org/10.1016/j.compedu.2019.05.017

87.

Wan

Xie

Shu

(2020). Toward an understanding of university students’ continued intention to use MOOCs: When UTAUT model meets TTF model. SAGE Open, 10(3), Article 215824402094185. https://doi.org/10.1177/2158244020941858

88.

Wang

(2024). Effective skills supply and grand vocational education: Interdisciplinary thinking on China's vocational education. Vocation, Technology & Education, 1. https://doi.org/10.54844/vte.2024.052

89.

Chen

(2017). Enhancing critical thinking skills through online collaborative learning: A study in higher education. Internet and Higher Education, 32, 37–45. https://doi.org/10.1016/j.iheduc.2016.06.002

90.

Xiang

Niu

(2022). Research on handicraft design based on artificial intelligence technology in complex environments. Wireless Communications and Mobile Computing, 2022(1), 1538488.

91.

(2023). Challenges and strategies of AIGC in the design industry. Artificial Intelligence, 2023(4), 3–4. https://doi.org/10.16453/j.2096-5036.2023.04.006

92.

Yang

Zhang

Qiu

Zhang

(2021). Intelligent design of product forms based on design cognitive dynamics and a cobweb structure. Computational Intelligence and Neuroscience, 2021(1), 1–17. https://doi.org/10.1155/2021/6654717

93.

Yener

Arslan

Kılınç

(2020). The moderating roles of technological self-efficacy and time management in the technostress and employee performance relationship through burnout. Information Technology and People, 34(7), 1890–1919. https://doi.org/10.1108/itp-09-2019-0462

94.

Yin

Zhang

Liu

(2023). The exploration of integrating the midjourney artificial intelligence generated content tool into design systems to direct designers towards future-oriented innovation. System, 11(12), 566. https://doi.org/10.3390/systems11120566

95.

Zaremohzzabieh

Roslan

Mohamad

Ismail

I. A.

Jalil

H. A.

Ahrari

(2022). Influencing factors in MOOCs adoption in higher education: A meta-analytic path analysis. Sustainability, 14(14), Article 8268. https://doi.org/10.3390/su14148268

96.

Zhang

Qiu

Zhang

(2022). Does technostress increase R&D employees’ knowledge hiding in the digital era? Frontiers in Psychology, 13, Article 873846. https://doi.org/10.3389/fpsyg.2022.873846

97.

Zhao

Zhang

(2023). Design of artificial intelligence cultural creative industry based on machine learning. Soft Computing, 1–12. https://doi.org/10.21203/rs.3.rs-2789470/v1

98.

Zhou

Xie

Liang

(2023). How to increase consumers’ continued use intention of artificial intelligence voice assistants? The role of anthropomorphic features. Electronic Markets, 33(1), Article 60. https://doi.org/10.1007/s12525-023-00681-0

99.

Zhou

Wang

(2010). Integrating TTF and UTAUT to explain mobile banking user adoption. Computers in Human Behavior, 26(4), 760–767. https://doi.org/10.1016/j.chb.2010.01.013

100.

Zigurs

Buckland

B. K.

(1998). A theory of task/technology fit and group support systems effectiveness. Mis Quarterly, 22(3), Article 313. https://doi.org/10.2307/249668

Revealing the Role of Technostress in the Continuous Intention of Using AIGC Among Designers: An Investigation Based on UTAUT and TTF Models

Abstract

Plain Language Summary

Keywords

Introduction

Theoretical Background and Hypotheses

AIGC

UTAUT Model

TTF Model

User Satisfaction

Technostress

Critical Perspectives on Technology Adoption Frameworks

Materials and Methods

Participants and Data Collection

Instruments

Data Analysis

Results

The Measurement Model

The Structural Model

Discussion

Performance Expectancy, and Social Influence have a Positive Effect on the Continuous Intention of Using AIGC

Effort Expectancy Does Not Significantly Positively Impact Continuous Intention of Using AIGC

The Influence of Both Task Characteristics and Technology Characteristics on Task-Technology Fit is Substantial and Positive

The Relationship Between the UTAUT and TTF Model is Mainly Reflected in the Influence of Task-Technology Fit on Performance Expectancy

Task-Technology Fit, In Turn, Also Exerts a Significant Positive Effect on User Satisfaction. User Satisfaction Positively Influences Designers’ Continuous Intention

Negative Moderating Effect of Technostress

Conclusion and Implications

Limitations and Future Research

Footnotes

Appendix

Acknowledgements

ORCID iDs

Ethical Considerations

Consent to Participate

Consent for Publication

Author Contributions

Data Availability Statement

References