Exploring Designer Trust in Artificial Intelligence-Generated Content: TAM/TPB Model Study

Abstract

Traditionally, users have perceived that only manual laborers or those in repetitive jobs would be subject to technological substitution. However, with the emergence of technologies like Midjourney, ChatGPT, and Notion AI, known as Artificial Intelligence-Generated Content (AIGC), we have come to realize that cognitive laborers, particularly creative designers, also face similar professional challenges. Yet, there has been relatively little research analyzing the acceptance and trust of artificial intelligence from the perspective of designers. This study integrates the TAM/TPB behavioral measurement model, incorporating intrinsic characteristics of designers, to delineate their perceived risks of AIGC into functional and emotional dimensions. It explores how these perceived characteristics, risks, and trust influence designers’ behavioral intentions, employing structural equation modeling for validation. The findings reveal the following: (1) designer trust is the primary factor influencing their behavioral choices; (2) different dimensions of perceived risks have varying degrees of impact on trust, with functional risks significantly positively affecting trust compared to emotional risks; (3) only by enhancing the transparency and credibility of Artificial Intelligence-Generated Content (AIGC) can the perceived characteristics of designers be elevated; and (4) only by effectively safeguarding designers’ legitimate rights and interests can perceived risks be significantly reduced, thereby enhancing trust and subsequently prompting actual behavioral intentions. This study not only enhances the applicability and suitability of AIGC across various industries but also provides evidence for the feasibility of intelligent design in the creative design industry, facilitating the transition of AIGC to Artificial Intelligence-Generated Design (AIGD) for industrial upgrading.

1. Introduction

The widespread adoption of Artificial Intelligence-Generated Content (AIGC) technology has become a hallmark of our era, acting as a double-edged sword for the design industry. Some designers worry that AIGC will lower industry standards, thereby affecting their income and career development. Conversely, others are eager to use AIGC to enhance work efficiency and broaden the scope of applications. As academic researchers, we are more concerned with the impact of AIGC technology on the design ecosystem: from design education and training at the front end, to design research and innovation and commercial practice in the mid-stage, and finally to design generation and evaluation at the back end. Therefore, this paper uses designers’ basic drawing skills as a starting point to explore their acceptance of AIGC. This study employs a combination of expert interviews and questionnaires. Independent sample T-tests and variance analysis were used for data analysis. The results indicate that designers with high professional recognition and relevant experience with AIGC exhibit a more positive attitude towards the technology. In contrast, those without relevant experience demonstrate a relatively negative attitude. Designers’ acceptance of AIGC is somewhat influenced by their existing drawing skills. These findings can provide valuable references for design education, design services, and other researchers.

AIGC has been widely applied in various aspects of the art and design field, accelerating design visualization and expanding design possibilities, thus fostering the flourishing development of the art and design industry [1]. Its impact on the field of art and design is mainly reflected in three areas: the internal working methods of the industry, the external social influence, and human–machine collaborative innovation. Specifically, it encompasses the following: (1) It has transformed the work content, processes, and design thinking of designers. It reduces the workload in the ideation phase, allowing designers to focus more on refining design definitions and deepening design thinking [2]. (2) It has lowered the industry threshold for design, enabling cross-disciplinary knowledge integration. By diminishing the skill barriers in the art and design industry, it has popularized innovation awareness across society, allowing innovators from various fields to leverage the efficiency of AIGC to collectively advance the art and design industry [3]. (3) Designers will achieve specialization and personalization of AI assistants through model training. Mastery of AIGC will become a fundamental skill for future designers, who will actively seek the optimal convergence of technology and human creativity, thereby attaining an efficient human–machine collaborative innovation working model [4]. In conclusion, AIGC, driven by data-driven innovation, has inspired an industrial revolution in the art and design field, optimizing design processes, methods, and quality, and thereby propelling the rapid development of the art and design industry.

Current academic research predominantly focuses on technology research and applications, with limited studies addressing users’ attitudes and behavioral changes towards artificial intelligence technology, especially in the field of art and design. This study aims to analyze designers’ acceptance and trust in AI-generated content (AIGC), exploring how designers’ intrinsic characteristics influence perceived risk (both functional and emotional) and trust. Additionally, it examines the impact of perceived characteristics, perceived risk, and trust on designers’ behavioral intentions. By enhancing the transparency and credibility of AIGC and safeguarding designers’ legitimate rights, this study aims to reduce perceived risk and increase trust, thereby promoting designers’ behavioral intentions towards AIGC. This research not only enhances the generality and compatibility of AIGC across various industries but also supports the intelligent transformation of the creative design industry, facilitating the upgrade from AI-generated content to Artificial Intelligence-Generated Design (AIGD).

In this study, the Technology Acceptance Model (TAM) and the Theory of Planned Behavior (TPB) were employed to explain changes in designers’ behavior. Previous behavioral models often explored the impact of trust or perceived risk on behavior change in isolation, merely confirming that trust in technology significantly positively influences its extended use, without considering the influence of perceived risk on usage behavior. Based on these classic theoretical foundations, this paper adopts an integrated model of TAMs and the TPB as the primary research framework. It primarily examines the hierarchical relationship between the antecedent factors in the TAM/TPB and perceived risk, as subdivided into specific risk dimensions, and their relationship with trust. Furthermore, it establishes a behavioral model that affects users’ willingness to use AIGC. This provides a theoretical basis for the widespread use and rapid development of AIGC in the future and offers theoretical support and practical guidance for AIGC technology providers to understand designers’ risk perceptions and implement risk mitigation measures.

The research results showed a significant gap from the expected outcomes, with only 4 out of the 11 proposed hypotheses being supported. This indicates that, in the field of art and design, designers hold a generally negative attitude towards AIGC. This finding highlights areas for improvement in the future development of artificial intelligence technology. To explore the key factors influencing designers’ behavior, our research focuses not merely on the binary judgment of users accepting or rejecting new technology but delves deeper into the underlying logic behind their acceptance or rejection of new technology. Our aim is to uncover the risk factors or trust elements that users consider when making decisions, as well as the reasons and motivations behind their attitudes towards new technology. By doing so, we can adjust the work methods and content of future designers to better meet users’ needs and expectations. Users’ attitudes towards new technology are influenced not only by personal preferences but also by a combination of factors. These factors may include users’ perceptions of the technology, its usability and ease of use, its potential impact on their daily lives, and the interaction of personal and social factors. By deeply researching these factors, we can better understand the process of attitude formation towards new technology. Analyzing the logic behind users’ acceptance or rejection of new technology can provide valuable insights for future designers. These insights can guide designers in considering users’ expectations and needs when developing new technologies, thereby better meeting users’ expectations. Additionally, by adjusting designers’ work methods and content, we can improve the acceptability and adoption rate of new technologies, further promoting technological advancement and societal progress.

In conclusion, this study delves into the underlying logic behind users’ acceptance or rejection of new technology and uses these findings to improve future designers’ work methods and content. This has a positive impact on enhancing collaborative innovation between designers and AIGC, not only increasing designers’ work efficiency but also optimizing their user experience when using new technology [5]. Its contributions are as follows: on one hand, from the perspective of technology providers, it offers users more trustworthy, user-friendly, and efficient products; on the other hand, from the perspective of user acceptance and understanding, it seeks the best collaborative innovation model with AIGC. We believe that such research will help promote the adoption and development of artificial intelligence technology, providing designers with more efficient functional value and safer emotional value.

2. Literature Review

AIGC has gained widespread recognition and application in both academic and commercial fields. On 22 August 2023, Harvard University’s Dean of the Faculty of Arts and Sciences, Christopher Stubbs, and Dean of Undergraduate Education, Amanda Claybaugh, hosted an informational meeting on “How to Use Generative AI in Courses Scientifically” and subsequently released the “AI Usage Guidelines and FAQs”. The guidelines clearly state that Harvard University supports the use of generative AI in teaching by faculty and students, provided the technology’s safety, privacy, regulatory compliance, and academic integrity are ensured [6]. Simultaneously, AIGC plays a crucial role in commercial environments by enhancing profitability through data analysis and performance metrics [7]. For instance, AIGC integrates machine learning and big data to improve the efficiency and accuracy of medical diagnoses [8]. In summary, the application scope of AIGC has extended from the commercial sector to various areas such as healthcare and education. The collaborative working methods facilitated by AIGC have garnered widespread public support.

2.1. Design Thinking

In the field of art design, design thinking has increasingly been integrated into the development of AIGC solutions [9]. From the perspective of research motivation and objectives, scholar Meng Zhang and others have proposed the effectiveness and application trends of combining AIGC with the art design field [10]. From the perspective of research significance and value, scholar Zaphir and others have introduced a framework for evaluating the quality of generative AI thinking [11]. From the perspective of research methods, scholar Hu Yin and others have suggested AI design tools and innovative methods. However, there is limited research on the principles and operational processes of design thinking in the context of AI technology [2]. Therefore, this study explores the impact of AIGC on the principles and processes of design thinking, as well as designers’ acceptance and trust, through empirical research and qualitative interviews. Figure 1 depicts the design thinking model revised by NN Group [12].

Additionally, a designer’s professional competencies can be summarized as external skills and internal design thinking. Design skills include standardized abilities such as sketching, software usage, and model making. Design thinking encompasses foundational theories, work processes, and methods. Originating from Stanford University D. school, design thinking is a systematic innovation process [13]. As its application has broadened, the concept has gradually extended to include definitions, considerations, and innovative perspectives on design [14]. This study focuses on designers’ perspectives, specifically within the realm of design thinking, exploring their attitudes towards and intentions of using AIGC.

However, the literature reveals that a significant factor in users’ refusal to use AIGC is the lack of trust [15]. Scholar Oscar Oviedo-Trespalacios explores methods of information acquisition from an ethical perspective to reduce users’ concerns about the security of AIGC information [16]. Scholar Monika Hengstler examines factors that enhance user trust in AIGC technology in commercial environments, specifically from the perspectives of healthcare and autonomous driving [17]. Scholar Martina Benvenuti investigates the implementation of personalized education through AIGC and intelligent assistance in primary education, thereby increasing the acceptance of new technology in school environments [18]. In other words, the current research focus on AIGC technology is on theoretical and practical measures to increase user trust.

Previous research on the applicability of AIGC often considers external environments, with limited focus on users’ personal traits and behaviors. Traditional designers are generally enthusiastic about using artificial intelligence to address design issues and show a high acceptance of AIGC technology [19]. However, as the scope of tasks assigned to AIGC technology expands, it may lead to designers’ feelings of insecurity. In other words, designers are more willing to accept AIGC technology as an assistant in completing design tasks rather than as a primary driver of the work. Therefore, it is crucial to consider how to balance the contributions and responsibilities of AIGC technology in design tasks. This study focuses on designers and explores the impact of perceived characteristics and trust on designers’ behavioral intentions from the perspective of perceived risk [20].

2.2. Theoretical Framework

2.2.1. Technology Acceptance Model

In 1989, scholar Davis introduced the Technology Acceptance Model (TAM) to explain and predict public acceptance of new technologies. This theory posits that users’ intention to use a technology is decisive for actual usage behavior and that attitudes towards the technology, along with perceptions of its usefulness and ease of use, jointly determine this intention. Specifically, attitude refers to the user’s subjective positive or negative feelings towards using the new technology [11]. Perceived usefulness is the belief that the technology can enhance work efficiency and quality, while perceived ease of use is the belief that the technology has low learning costs. Using the framework provided by TAMs, researchers can easily understand users’ attitudes towards new technology and predict their usage behavior. This theory has been widely applied across various high-tech fields, including the internet, smart manufacturing, and AIGC collaboration [21].

The original framework of the Technology Acceptance Model (TAM) and its extended models are frequently used to understand users’ acceptance of new technologies. They measure perceived usefulness, perceived ease of use, and usage intention and predict usage behavior. In the context of the artificial intelligence era, TAM’s applicability is even broader. For instance, scholar Holdack used TAMs to understand users’ acceptance of AR wearable devices, finding that perceived enjoyment and information richness are key factors influencing behavior [22]. Scholar Zhang applied TAMs to explore users’ acceptance of autonomous vehicles, identifying social influence and initial trust as critical factors affecting technology use [23]. Scholar Vahdat employed TAMs to investigate customers’ purchase intentions for mobile application technologies, concluding that social factors and peer influence are crucial determinants of purchasing behavior [24].

Overall, a TAM proves to be universally applicable across various domains. Additionally, extended TAMs have received broad recognition. For example, scholar Chiu combined TAMs with the Theory of Reasoned Action (TRA) to form the TRAM framework, exploring the role of technological readiness in individuals’ intentions to use health products and fitness applications [25]. Scholar Al-Maatouk integrated a TAM with the Task–Technology Fit (TTF) model to create a new theoretical framework, examining students’ satisfaction with learning social media [26]. Furthermore, scholar Kamdjoug merged a TAM with the UTAUT2 model to measure users’ attitudes towards mobile banking applications and the factors affecting behavioral change. These studies confirm that a TAM and its extended models are effective in understanding users’ acceptance of various technologies [22]..

2.2.2. Theory of Planned Behavior

In 1977 and 1980, scholars Ajzen and Fishbein introduced the Theory of Reasoned Action (TRA), which posits that users’ intentions determine their behavior and can be evaluated and measured [18]. To expand and enrich the application of this model, scholar Icek Ajzen developed the Theory of Planned Behavior (TPB) in 1988 and 1991 [27]. The TPB is used to explain users’ intentions and behavioral changes by deconstructing the factors related to users’ intentions or behavior changes. Building on the TRA, the TPB incorporates the evaluation standard of perceived behavioral control, allowing it to explain the logic behind involuntary behavior changes and eventually evolve into the TPB.

The TPB comprises three evaluative criteria: attitude, subjective norms, and perceived behavioral control. Attitude refers to the user’s evaluation of a specific behavior, subjective norms denote the influence of others’ expectations and social pressure on the user, and perceived behavioral control represents the user’s perceived control over the behavior. These three combined factors ultimately affect users’ behavioral intentions. The theory is widely applied in explaining individual users’ behavior and predicting group behavior, providing robust theoretical support for the application of new technologies.

The Theory of Planned Behavior (TPB) is a widely used social psychological model that not only explains behavioral changes but also predicts future behaviors. The model posits that users’ behavioral intentions are the direct factors determining their actual behavior. The key factors influencing users’ behavioral intentions are as follows: (1) behavioral attitude—the user’s positive or negative attitude towards performing a particular behavior; (2) subjective norms—the perceived social expectations and pressures related to the behavior; and (3) perceived behavioral control—the user’s perception of the difficulty of performing the behavior.

The TPB model has been successfully applied in various fields. For example, scholar Blanco-Mesa used the TPB to explore the factors influencing entrepreneurial intentions among Colombian university students, finding that personal attitude and perceived behavioral control are direct variables affecting entrepreneurial intentions, while subjective norms act as an indirect variable [28]. Scholar Yuriev applied the TPB to study users’ environmental behaviors, concluding that user beliefs are an indirect variable influencing behavior [29]. Scholar Erul used the TPB to explain users’ intentions to support tourism development, finding that emotional factors are key to behavioral change [30]. Scholar Gieure applied the TPB to explore the link between entrepreneurial intentions and behavior, identifying self-efficacy as an influential factor in entrepreneurial activities [31]. Additionally, several scholars have extended the TPB; for instance, Kim combined the Norm Activation Model with the TPB to explain pro-environmental behavior in the context of drone delivery services, while Ha integrated the TPB with a TAM to investigate the impact of perceived risk on consumers’ online shopping intentions [32]. Overall, the TPB provides a robust framework for understanding user attitudes and behavioral intentions, offering valuable support for predicting behavioral changes.

2.2.3. The Theoretical Model of This Study: Integration of TAMs and the TPB

Although the integration of the Technology Acceptance Model (TAM) and the Theory of Planned Behavior (TPB) is widely applied in research on user behavior change and technology adoption, studies on its application in the field of art and design within the context of artificial intelligence are relatively scarce. This scarcity can be attributed to two main reasons: On one hand, the rapid adoption of artificial intelligence technology across various domains has yet to yield a systematic and authoritative academic discourse. On the other hand, the art and design field, characterized by both rational and emotional aspects, demands innovation and uniqueness, which contrasts sharply with the standardized solutions offered by AIGC. This discrepancy highlights the challenges artificial intelligence will face in expanding its applications. Therefore, it is essential to propose new research models to explore the potential applications of artificial intelligence in the field of art and design.

The integration of the Technology Acceptance Model (TAM) and the Theory of Planned Behavior (TPB) is commonly used to understand consumers’ attitudes and behavioral changes towards new technologies, including AI-based intelligent products [33]. This integrated model has established precedents across various fields. For instance, in traditional agriculture, scholar Mohr used TAMs and the TPB to study German farmers’ acceptance of AI and the factors influencing their behavior [34]. In e-commerce, it has measured elderly people’s continued willingness to engage in online shopping [35]. In higher education, it has explored students’ acceptance of MOOCs and behavioral changes [36]. In healthcare, it has assessed the impact of COVID-19 on users’ shopping behavior [37]. Scholar Blanco-Mesa integrates the Theory of Planned Behavior (TPB) with Partial Least Squares Path Modeling (PLS-PM) in structural equation modeling (SEM), identifying personal attitude (AC) and perceived behavioral control (CCP) as key influencing factors in users’ behavioral changes [28]. These studies collectively affirm the broad applicability of TAMs and the TPB, showing how these frameworks can help understand the factors affecting designers’ attitudes and behavioral changes towards AI technology.

A TAM focuses on two subjective variables—perceived usefulness and perceived ease of use—and their effects on user intentions and behaviors, while the TPB addresses two feedback variables influenced by external factors: subjective norms and perceived behavioral control. Integrating these models can further elucidate users’ attitudes towards new technologies and predict behavioral changes. Many scholars have applied this combined model in various fields to enhance its explanatory power. By merging TAMs with the TPB, the limitations of each individual model are addressed, providing a comprehensive evaluation framework that includes perceived usefulness, perceived ease of use, subjective norms, and perceived behavioral control [38]. Consequently, this research, based on the integration of TAMs and the TPB, incorporates direct variables of functional and emotional risks, effectively explaining the key factors influencing designers’ attitudes [39].

This study integrates the TAM/TPB research measurement model to explore the impact of three influencing factors—perceived characteristics, perceived risk, and trust—on designers’ behavior, as well as the logical relationships among the antecedent factors. The structural equation model is employed to test the validity of the experimental hypotheses. Based on the original TAM/TPB framework, designers’ perceived characteristics include perceived usability, perceived ease of use, and subjective norms. Considering designers’ intrinsic traits, perceived risk is divided into functional risk and emotional risk. Trust refers to designers’ willingness to use AIGC, which directly determines their actual usage behavior.

In the context of the AI era, integrating TAMs and the TPB effectively explains the applicability of AIGC in the field of art and design, as well as understanding designers’ attitudes and usage intentions towards AIGC. Thus, this research proposes a conceptual model, as illustrated in Figure 1, based on TAMs and the TPB, comprising 7 variables and 11 research hypotheses. The proposed model translates key variables from TAMs and the TPB into factors influencing designers’ use of AIGC, ultimately identifying the critical variables affecting designers’ adoption of AIGC.

2.3. Basic Hypotheses

2.3.1. Trust of Designers and Intention to Use AIGC

Research indicates that trust positively influences user behavior [40]. Trust is a key factor in converting users’ behavioral intentions [41]. Trust enhances designers’ perception of new technologies, reducing learning costs and increasing interest in learning [42]. Based on previous research findings, the following hypotheses are proposed:

H1. 

The trust of designers will have a positive impact on the behavior, behavioral intention, of AIGC.

2.3.2. Perceived Risks and Trust of Designers

Research indicates that users’ perceived risk can be categorized into two dimensions: functional and emotional risk [35]. This aligns with the characteristics of the art and design field, where design skills are divided into external design expression and internal design thinking abilities. When external technological environments change, designers experience functional and emotional risks, which in turn influence their attitudes towards technological transformations.

Perceived risk, as a fundamental characteristic of trust, also affects designers’ behavior. This correlation primarily manifests in how perceived risk diminishes trust [43]. Based on previous research findings, the following two hypotheses are proposed:

H2a. 

Functional risks will have a negative impact on the trust of designers.

H2b. 

Emotional risks will have a negative impact on the trust of designers.

2.3.3. Designers’ Perceptions and Trust Based on TAM/TPB Theory

AIGC technology enhances industrial structure and work efficiency, demonstrating significant market potential. When users face risks associated with new technology, their subjective norms, perceived usefulness, and perceived ease of use will influence their perception and valuation of AIGC, thereby affecting designers’ trust and intention to use AIGC. According to the definitions in TAM and TPB theories, perceived usefulness refers to the extent to which designers believe AIGC can improve their work efficiency, reflecting the technology’s impact on productivity. Perceived ease of use denotes how easy designers find it to use AIGC, indicating the ease with which they can learn and utilize the technology. Subjective norms refer to the consideration of opinions from other users or social groups before using AIGC. Thus, designers’ perceptions and trust are interrelated; perceived usefulness, ease of use, and subjective norms all positively influence trust [44]. Based on previous research findings, the following three hypotheses are proposed:

H3. 

Perceived usefulness has a positive impact on designers’ trust.

H4. 

Perceived ease of use has a positive impact on designers’ trust.

H5. 

Subjective norms have a positive impact on designers’ trust.

2.3.4. Designers’ Perceptions and Perceived Risks

Traditional research often treats designers’ perceived characteristics and perceived risks as common antecedent factors, studying their joint impact on the intention to use AIGC. In real environments, designers often make subjective assessments of perceived risks based on their self-perceptions [43]. It is evident that perceived usefulness and ease of use of new technologies will have a significant impact on designers’ willingness to use them, but the extent and manner of their influence have not been clearly studied. Based on previous research findings, the following four research hypotheses are proposed:

H6a. 

Perceived usefulness has a negative impact on functional risk.

H6b. 

Perceived usefulness has a negative impact on emotional risk.

H7a. 

Perceived ease of use has a negative impact on functional risk.

H7b. 

Perceived ease of use has a negative impact on emotional risk.

2.3.5. Designers’ Subjective Norms and Functional Risks

The theories of the TPB, TRA, and TRI emphasize the importance of subjective norms in individual behavior change [45]. Research has shown that subjective norms significantly influence behavioral intentions, surpassing attitudes, perceived usefulness, and perceived ease of use in explaining behavior change [46]. In the field of consumer behavior, subjective norms explain factors influencing individual consumption behavior [47]. In business, they explain the influence of entrepreneurs’ personal characteristics on business behavior [48]. In the realm of mobile payments, they explain the environmental impact on individual behavior [49]. In summary, subjective norms influence the intentions and actual behavior of individuals in various domains. Based on previous research findings, the following two research hypotheses are proposed:

H8a. 

Subjective norms have a negative impact on functional risk.

H8b. 

Subjective norms have a negative impact on emotional risk.

3. Research Methodology

3.1. Research Participants

To ensure the representativeness and typicality of the study participants, all respondents were professionals with a background in product design. They had completed systematic courses in design theory, possessed skills in hand drawing and computer graphics, and were capable of independently completing design projects, meeting the level of junior designers. In the analysis of the questionnaire, the final 92 valid samples were characterized as follows: (1) gender distribution: 120 females (34.8%) and 64 males (65.2%); (2) age distribution: 32 participants were 20 years old (32.6%), 34 participants were 21 years old (36.9%), and 28 participants were 24 years old (30.4%); (3) education level: 80 undergraduates (87%) and 12 master’s students (13%); and (4) AIGC product usage experience: 75% of respondents reported having used AIGC and being proficient in its methods and processes, while 25% expressed interest in AIGC and planned to use the technology in future design projects. Specific demographic information is shown in

Table 1. Basic information of interviewees (N = 184).

Profile	Items	Number	Percentage (%)
Gender	Male	64	34.8%
	Female	120	65.2%
Age	20	60	32.6%
	21	68	36.9%
	24	56	30.4%
Education	Undergrad	160	87%
	Postgraduate	24	13%
Major	Product design	92	100%
AI tools usage experience	Yes	138	75%
	No	46	25%

.

3.2. Questionnaire Design

To ensure the reliability and validity of the questionnaire, the items were derived from established scales. The questionnaire includes sections on personal basic information, perceived usefulness, perceived ease of use, subjective norms, trust, behavioral intention, functional risk, and emotional risk. The details of the questionnaire items and their sources are listed in

Table 2. Variable construction and sources.

Variables	Construct (Sources)	Question
Perceived usefulness	[50]	PU1. Using Midjourney makes my work more convenient.
		PU2. Using Midjourney allows me to get some useful information.
		PU3. Using Midjourney enriches my life and work.
Perceived ease of use	[27,50,51]	PEOU1 I think the operation of the Midjourney is simple.
		PEOU2 I think learning to use Midjourney is easy.
		PEOU3 I find it easy to use Midjourney.
Subjective norm	[50]	SN1 My customers thinks I should use Midjourney.
		SN2 My teachers & friends think I should use Midjourney.
		SN3 My friends and colleagues think I should use Midjourney.
Trust	[50]	TRU1 AIGC is trustworthy.
		TRU 2 AIGC values the interests of designers.
		TRU 3 AIGC is safer and more reliable than other design tools.
Behavioral intention	[40,45,52]	BI1. I am interested in using AIGC.
		BI2 I am willing to learn and use more features developed by AIGC.
		BI3 I am willing to increase the frequency of using AIGC
Functional Risk		FR1 AIGC protects user information security.
		FR2 AIGC authenticates the user’s identity.
		FR3 AIGC can ensure that information is transmitted without being tampered with or lost.
		FR4 If I use AIGC, my customers, teachers, friends and colleagues may think negatively about my behaviour.
Emotional Risk	[40,45,52]	ER1 AIGC is a new type of product, and using it creates unnecessary tension.
		ER2 Feeling uncomfortable when using AIGC.
		ER3 Psychological stress if financial loss occurs.
		ER4 When using AIGC, my personal information may be disclosed.

. The structured questionnaire covers 7 factors and 22 measurement items. It employs a five-point Likert scale, where a response of 1 indicates “strongly disagree” and a response of 5 indicates “strongly agree”. The completed questionnaire will be reviewed and refined by scholars and experts. The survey targets undergraduates and postgraduates with a background in design.

3.3. Data Collection

As I am teaching a course on “AI-Assisted Design”, the experimental design was integrated into the course exercises. The study employed a combination of semi-structured questionnaires and expert interviews. The questionnaires were distributed both online and offline. The response rate reached 85%, with a total of 198 questionnaires collected. Out of these, 14 were invalid, resulting in 184 valid responses.

4. Research Results

4.1. Reliability and Validity Testing of the Measurement Model

First, the suitability of the collected data was examined. Using SPSS Statistics 24.0, the six dimensions describing perceived risk were validated, with results indicating that the data are appropriate for this study. The KMO measure was 0.969, and the significance probability of Bartlett’s sphericity test was 0.000, demonstrating that the data fully meet the feasibility criteria for principal component analysis.

Next, the reliability of the model was tested using Cronbach’s alpha, a widely accepted method in academia. The results showed that the Cronbach’s alpha coefficients for the seven dimensions of the model—perceived usefulness, perceived ease of use, subjective norms, functional risk, emotional risk, trust, and intention—were all above 0.8, indicating high reliability of the questionnaire.

Finally, the convergent validity of the model was tested using confirmatory factor analysis (CFA). The measurement model’s fit must meet acceptable standards before further evaluation of the structural model. As shown in Table 2, all standardized factor loadings were above 0.7; composite reliability was higher than 0.8; and average variance extracted (AVE) exceeded 0.5. Thus, the model demonstrates strong convergent validity. The specific statistical data are shown in

Table 3. Confirmatory factor analysis.

Latent Variables	Observed Variables	Standardized Factor Loading	Cronbach’s α	CR (Critical Ration)	AVE
Perceived usefulness	PU1 PU2 PU3	0.960 0.962 0.953	0.978	0.9712	0.9184
Perceived ease of use	PEOU1 PEOU2 PEOU3	0.954 0.973 0.974	0.988	0.9774	0.9352
Trust	TRU1 TRU2 TRU3	0.924 0.968 0.912	0.954	0.9542	0.8742
Behavioral intention	BI1 BI2 BI3	0.961 0.793 0.868	0.937	0.9083	0.7686
Emotional risk	ER 1 ER 2 ER 3 ER 4	0.662 0.711 0.899 0.508	0.748	0.7952	0.5025
Subjective norm	SN1 SN2 SN3	0.650 0.567 0.753	0.705	0.7975	0.5012
Functional risk	FR1 FR 2 FR 3 FR 4	0.746 0.439 0.783 0.802	0.806	0.7937	0.5014

.

4.2. Structural Equation Model-Based Analysis

4.2.1. Overall Fit Evaluation of Structural Equation Model

The data were imported into SPSS AMOS 21.0 to explore the relationships between variables.

Table 4. Fit index values of the structural equation model.

Adaptation Index	Recommended Value	Fitted Value
χ2	The smaller the better	484.006
χ2/df	<3.0	2.251
GFI	>0.9	0.743
AGFI	>0.8	0.871
RMSEA	<0.08	0.078
NNFI	>0.9	0.901
IFI	<0.9	0.928
CFI	>0.9	0.903

lists the key fit indices obtained from the structural model assessment. Upon comparing the calculated values with the recommended fit indices, it is evident that all indices fall within acceptable ranges except for the GFI value, which is slightly below the recommended value of 0.743. Overall, the structural equation model’s setup is deemed acceptable. The model fit indices are presented in Table 4.

4.2.2. Results of Hypothesis Testing

The structural relationships between latent variables and the estimated values of standardized path coefficients were computed using SPSS AMOS 21.0, as detailed in

Table 5. Hypothesis validation results.

Code	Path	β	T-Value	p-Value	Outcome
H1	TRU → BI	0.118	7.294	0.000	Supported
H2a	FR → TRU	0.138	2.781	0.047	Supported
H2b	ER → TRU	0.201	0.019	0.985	Unsupported
H3	PU → TRU	0.120	2.675	0.007	Unsupported
H4	PEOU → TRU	0.097	−1.172	0.241	Unsupported
H5	SN → TRU	0.099	1.717	0.086	Unsupported
H6a	PU → FR	0.111	1.005	0.315	Unsupported
H6b	PU → ER	0.110	3.456	0.000	Supported
H7a	PEOU → FR	0.100	4.863	0.000	Supported
H7b	PEOU → ER	0.080	−0.289	0.772	Unsupported
H8a	SN → FR	0.122	1.101	0.271	Unsupported
H8b	SN → ER	0.116	0.889	0.374	Unsupported

. The actual model and path coefficients obtained are depicted in Figure 2.

In this study, the significance of path coefficients was examined using the bootstrapping method, and the results are presented in Table 4. Out of the 12 hypotheses proposed, 4 were validated, while 8 hypotheses were not supported, as illustrated in Figure 2.

Specifically, perceived ease of use significantly influences functional risk (β = 0.1, t = 4.863, p < 0.05), thus supporting hypothesis H7a. Functional risk significantly affects trust (β = 0.13, t = 2.789, p < 0.05), confirming hypothesis H2a. Trust significantly influences behavioral intention (β = 0.118, t = 7.294, p < 0.05), validating hypothesis H1. Furthermore, perceived usefulness significantly impacts emotional risk (β = 0.11, t = 3.456, p < 0.05), thereby supporting hypothesis H6a. It is evident that when designers confront the impact of new technology on the creative industry, they still weigh the acceptance of new technology from rational, functional, and risk-control perspectives.

The reasons why emotional risk and subjective norm do not significantly influence behavioral intention lie in users’ lack of trust in new technology, and hesitance to change established workflow habits. Perceived usefulness and perceived ease of use fail to significantly impact behavioral intention because most users already have a preliminary understanding of AIGC, acknowledging its efficient work methods, and regarding its functional aspects as basic rather than differentiating features that attract them to use it. The lack of significant influence of perceived usefulness and subjective norm on functional risk is attributed to users’ reluctance to adopt new technology, primarily stemming from external factors such as legal regulations, rather than being determined solely by subjective perceptions and experiences. Similarly, perceived ease of use and subjective norm fail to significantly influence emotional risk due to users’ insufficient experience with new technology and the high learning costs involved. It is evident that during the early stages of AIGC development, external factors such as legal regulations and individual user experience will influence its scope of application and frequency of use. The results of the research hypothesis are shown in Figure 3.

To gain further insights into the motivations behind designers’ usage patterns, researchers randomly selected 15 interviewees for in-depth interviews. The interviews focused on the usage experience and functional improvements of AIGC. The specific interview questions and responses from the interviewees are summarized in

Table 6. Main responses of semi-structured interviews.

Questions and Answers
Q1: What problems have you encountered in using AIGC or AIGC design tools?
“It can perform simple tasks, but not at a professional level.” (Interviewee 3, male).
“I need to keep reminding it of the requirements of the task in order for the AIGC to understand my needs.” (Interviewee 4, female).
“It works as a result of instability and is unable to perform different tasks in a stable and consistent manner.” (Interviewee 7, male).
Q2: What kind of AIGC design tool do you like?
“I like to use chatgpt to help me with research and analysis and then use midjourney for visual output tasks.” (Interviewee 1, male).
“I like to use Notion AI to help me summarise the design keywords and then use Stable Diffusion or midjourney to represent the preconceived image.” (Interviewee 2, female).
Q3: Why don’t you like working with AIGC? “I’m more concerned about the potential legal risks posed by the AIGC than its efficiency.” (Interviewee 10, female). “I’m afraid of becoming too dependent on AIGC’s capabilities and losing the drive to learn and grow myself.” (Interviewee 13, male). “In order to maintain a designer’s personality, I don’t like to upload my way of thinking and working to the web.” (Interviewee 15, female).

. Based on representative responses, three key factors influencing designers’ usage of AIGC were identified:

(1)

Incomplete Functionality of AIGC: Interviewee 3 and Interviewee 4, for instance, believed that AIGC is better suited as a design assistant because they require clear guidance from users to leverage the advantages of big data.

(2)

Professional Risks Associated with AIGC: Interviewee 13 and Interviewee 15 expressed concerns that AIGC could deprive them of their individuality and creativity, turning them into mere servants of technology, thus affecting their professional growth.

(3)

Incomplete Usage Environment of AIGC: Interviewee 10 highlighted concerns about intellectual property rights associated with AIGC, indicating that legal risks could make designers skeptical about new technology.

In conclusion, the current obstacles hindering designers’ acceptance of AIGC primarily stem from two aspects. Firstly, there are intrinsic factors related to the new technology itself, such as high learning costs, technological immaturity, unstable output content, and the need for explicit instructions. Secondly, external environmental factors, including risks related to intellectual property rights, the homogenization of big data, and incomplete legal frameworks, also contribute to the reluctance of designers to adopt AIGC. The above discussion primarily focuses on AIGC’s usage experience and functional improvements. Most respondents acknowledged its functional advantages but expressed a lack of trust in artificial intelligence technology. Specific concerns include legal risks, fairness, and worries about career development. Representative responses are detailed in Table 6.

In the context of artificial intelligence, design thinking will place greater emphasis on the designer’s ability to define market needs. The principles of design thinking encompass the following:

• Originality in Design Definition: Before creation, it is necessary to establish the working principles and processes of artificial intelligence through the designer’s work experience and intuitive judgment.
• High Efficiency of Design Proposals: During the creation process, actively utilize AIGC to generate a large number of renderings, thereby increasing rapid iteration and fault tolerance.
• Achieving AIGD (Artificial Intelligence-Generated Design) through Collaborative Innovation: At the end of the creation process, summarize the gains and losses to derive innovative principles and methods.

In other words, within the context of AI technology, designers can accurately grasp market and user needs through “Text Generative AI” and then visualize creativity with the help of “Image/Video Generative AI”. AIGC significantly shortens the application process of traditional design thinking, meeting ever-changing market demands through precise and rapid iterations.

In conclusion, during the AI-assisted design process, the research direction, clarity of the design definition, and the selection of design schemes need to be set by the designer based on their experience. However, AI technology can achieve higher work quality in research content, sketch diversification, the rendering of different design schemes, and the refinement of design details. This integration of human intuitive creativity and AI’s rational data represents an innovative process that will profoundly impact the future principles and processes of designers’ work. The innovative principles of AI design thinking emphasize a user-centered approach, proposing effective solutions through big data definition and rapid iteration by AIGC. The principles and processes of design thinking in the context of AI are illustrated in Figure 4.

5. Conclusions

This study’s findings indicate that with the widespread application of AIGC, users have developed a high sense of usability, with usability being a key determinant of whether AI products are adopted. In the creative design industry, the more prominent designers’ perceived characteristics are, the more effectively they can reduce perceived risks, deepen their trust in AIGC, and consequently increase their desire to use it. A detailed analysis is as follows:

(1)

The functional usability of AIGC positively influences emotional risk, implying that the feasibility of new technology can more easily affect users’ acceptance from a subjective emotional perspective. Designer trust is a key factor influencing behavioral judgment, while perceived risks negatively affect designer trust and behavior. Therefore, improving the usability of AIGC can enhance designers’ trust in it, thereby increasing its usage.

(2)

The functional usability of AIGC positively influences functional risk, indicating that the advantages of new technology over traditional ones more easily influence users’ choices from an objective acceptance perspective.

(3)

Users’ concerns about functional risk positively influence their intention to use, which in turn positively influences their actual usage behavior. Thus, the primary factor influencing the actual usage of AIGC is users’ consideration of whether its functionality is adequate.

In conclusion, designers are more concerned with the usability of AIGC to reduce their concerns about functional risk, thereby directly determining their intention to use and behavior. Furthermore, AIGC not only optimizes design efficiency but also significantly impacts design processes and workflows, which will be a focus of future research.

Based on the above research, we anticipate that AIGC will impact designers’ work methods in the following ways:

(1)

Transformation of Workflow: Designers will experience a significant shift from merely executing tasks to becoming guides. This change underscores the need for designers to grasp macro design directions and accurately define micro design keywords. Future designers will no longer just carry out assigned tasks but will assume the role of project leaders and decision-makers, offering strategic design guidance.

(2)

Evolution of Roles: Designers will evolve from being mere “creators” to also taking on the role of “supervisors”. AI technology will handle some design tasks automatically, altering the designer’s role to include reviewing and adjusting AI outputs.

(3)

Enhanced Innovation: Designers will need stronger innovation capabilities. As AI technology expands the feasibility and scope of design, designers must cultivate enhanced creativity and awareness to produce more innovative and forward-thinking design solutions.

(4)

Increased Emphasis on Collaboration: Given that AI technology involves expertise from multiple fields, designers will need to focus more on team collaboration and communication to effectively integrate diverse knowledge and resources, achieving better design outcomes.

In summary, despite many designers expressing concerns about AI technology, its potential to disrupt work methods, processes, and content is undeniable. Designers are preparing proactively to address these risks. AIGC is profoundly affecting industry development; while AI-generated content offers efficiency and personalization advantages, it also raises concerns about creativity, intellectual property, and content authenticity. As technology progresses, it is crucial to thoughtfully and ethically navigate the opportunities and challenges posed by AI-generated content.

6. Discussion

The impact of AIGC on designers’ workflows and tasks is evident in the following aspects: (1) Rapid Prototyping: AI technology, through machine learning algorithms, can automatically generate multiple design options, accelerating the prototyping process and enhancing designers’ efficiency; (2) Enhanced Design Quality: AI provides more accurate data and analysis, aiding designers in creating superior design solutions. It can also continuously optimize designs based on user feedback and historical data; (3) Increased Efficiency and Creativity: AI can produce a large volume of design proposals in a short time. Although these proposals are combinations of existing elements, they can inspire designers to engage in deeper thinking; (4) Data-Driven Design: By analyzing vast amounts of data, AI helps designers better understand user needs and market trends, leading to more informed design and decision-making. In summary, AIGC has gradually replaced traditional design tools, becoming an essential skill for designers.

6.1. Research Limitations

This study has provided an initial exploration of designers’ attitudes towards AIGC in the creative design field. However, there are several limitations that need to be addressed:

(1)

Time constraints: The experiment was conducted in early 2023 when AIGC products like ChatGPT were still in the early stages of promotion. Users were in the training phase for new technology, and designers had limited experience with it. It is believed that as the products are promoted further, the acceptance of AIGC will change.

(2)

Participant limitations: The majority (85%) of participants were students, who face less competition pressure compared to employed designers. Hence, their attitudes towards learning and accepting new technology may be more negative. Future research could separately analyze students and employed designers to understand the impact of usage environment and motivation on usage behavior.

(3)

Environmental constraints: With the widespread use of AIGC, corresponding changes in laws and learning pathways are expected, reducing users’ concerns about safety. Future research could incorporate variables related to the social environment into the model to form a more comprehensive measurement theory.

The limitations of this study primarily stem from its research background and scope. On one hand, the experimental design was conducted in September 2023, a period when users still harbored distrust and unfamiliarity with artificial intelligence. This reflects the gradual acceptance process of new technologies, which means the experimental results may not robustly support the hypotheses. As AI technology becomes more widespread, users’ habits with AIGC are likely to evolve, leading to more in-depth research in the future.

On the other hand, the study’s scope is restricted to designers’ acceptance and readiness for AIGC, without exploring the adaptability of AIGC to different tasks at a micro level. It lacks consideration of specific work contexts and objectives. Moreover, factors influencing designers’ tool choices, such as AI product usage experience, the cost of learning new tools, and the complexity of tasks, will be crucial research areas moving forward. Additionally, the ethical and legal risks associated with AIGC remain significant barriers to its rapid market adoption.

6.2. Research Trends

Currently, AIGC can explore innovations by combining various directions and styles, but the greatest strength of designers lies in their logic and uniqueness. While AIGC excels in creative divergence, designers are more capable of achieving transformative innovation. AIGC’s major limitation is that it cannot truly replace designers. Designers’ thinking is based on multidimensional, multifaceted, and interdisciplinary knowledge, driven by subjective creativity, whereas AIGC generates design solutions through combinatorial knowledge bases without reflecting subjective thought. In other words, AI diminishes the differentiation in designers’ external skills while inadvertently reinforcing the dominant role of design thinking. This shift brings the art and design industry back to the core focus on uncovering and defining user needs, guiding AIGC’s workflows and design proposals through design thinking, and ultimately achieving human–machine collaborative innovation. This has inspiring implications for design education, training, and commercial practice.

However, AI technology continues to expand and deepen its applications in the field of art and design. It offers designers more tools and possibilities while promoting innovation and development in design. Specific trends include the following:

(1)

AIGC as a more intelligent design tool: As AI technology advances, increasingly intelligent design tools will be developed. These tools will help designers complete tasks more quickly and accurately while providing intelligent optimization based on user feedback.

(2)

Enhanced integration and innovation across disciplines: AI’s need for expertise from multiple fields will foster interdisciplinary collaboration, leading to better design outcomes.

(3)

Development of design ethics and regulations: Governments or industry associations are expected to introduce design ethics and legal frameworks related to AIGC, ensuring fairness and transparency in AI algorithms while mitigating biases and discrimination.

(4)

Personalized and customized services: AIGC technology can customize design solutions according to a user’s preferences and needs, thereby enhancing user satisfaction.

(5)

Highly efficient and minimalist working methods: AIGC technology can simplify design processes and elements through automation and intelligence, improving design efficiency and conciseness.

(6)

Enhanced human–machine interaction experience: AIGC technology can transform design content into interactive experiences, increasing user engagement and experience satisfaction.

(7)

Future designers will focus on constructing systematic design thinking, as this will determine how they use AIGC to enhance their efficiency and work quality. At the same time, the principles of design thinking will shift from the role of “analyzing problems” to “posing questions”.

In summary, proficiency in applying AI technology to professional fields will become a fundamental skill for future designers. The design process will undergo significant changes, transitioning from a role of execution to one of guidance, with AIGC becoming an essential assistant and partner in the designer’s workflow.