Introduction

With rapidly enhanced information technology capabilities, extensive big data utilization, and growing public engagement, artificial intelligence (AI) has attracted significant attention1. Evidence from global employee surveys suggests that between 2018 and 2025, AI will play a pivotal role in catalyzing business expansion2. Currently, AI-generated content (AIGC) is being used progressively in a variety of sectors, such as the creation of text, audio, visual, and strategic content3, heralding a potential renaissance in content productivity. AI interventions have affected a significant number of tasks performed by American workers4. Scholars’ interpretations of well-known paintings have also been impacted by AI as an auxiliary tool5,6. AIGC has revolutionized the design field by broadening designers’ conceptual horizons and transforming creative methods and expressive techniques. For instance, AI painting tools can produce accurate visual representations from textual descriptions in approximately one minute, exemplifying the synergy between human creativity and machine efficiency. This innovation disrupts traditional practices and furnishes content creators with a more immediate conduit to their audiences7. Given these changes, proficiency in AI painting tools has become an essential skill for modern designers. Prior to the advent of AI painting tools, major corporations began employing AI to generate promotional materials such as posters and videos8. To facilitate the adoption of AI painting tools by designers, technology developers must comprehend the factors that influence designers’ intentions to use such tools. A thorough understanding of these determinants can help identify the barriers that prevent widespread adoption and provide guidance for formulating strategies that promote the incorporation of AI painting tools into the design procedure.

A review of the extant literature on AI painting tools reveals certain limitations. Studies specifically addressing designers’ inclinations to transition to AI painting tools are scarce. Existing research has primarily investigated whether AI-generated artworks possesses intrinsic artistic value9, whether the AI-generated image is similar to our interpretation of the same painting5,6, the effects of AI on artistic creation in the fields of painting and design10,11, and the ability of users to distinguish AI-generated images from those created by humans12. Furthermore, most research on AI painting tools uses the extended technology acceptance model (TAM) to explain user acceptance behavior8,13,14,15, which results in theoretical underpinnings that are frequently narrow in scope. This model fails to comprehensively encompass the complex range of motivating and hindering elements that affect designers’ willingness to switch to AI painting tools.

To address this issue, this study applies a push–pull–mooring (PPM) framework to the designers’ intentions to switch to AI painting tools. The PPM framework is used as a research method because it has significant advantages. The structure of the PPM framework is relatively simple, easy to understand and operate, and is supported by extensive empirical research. The PPM has been used to understand individuals’ intentions to switch between different technologies16. From an academic perspective, the adoption of AI painting tools by designers can be likened to a form of technological migration in which designers transition from traditional methods to novel technological paradigms. The PPM framework is appropriate for analyzing designers’ switching behaviors. The “push” factors may include the limitations or dissatisfaction with traditional painting tools, while the “pull” factors could encompass advanced features and benefits offered by AI painting tools. “Mooring” factors refer to a designer’s individual characteristics that may either facilitate or hinder this transition.

This study’s primary objective is to offer a comprehensively investigate the variables that influence designers’ decisions to move from conventional to AI painting tools. By concentrating on the creative design context, this investigation deeply explores the phenomenon of switching behavior in terms of three dimensions: the switching costs involved, the attractiveness of AI painting tools, and the dissatisfaction with traditional painting tools. Specifically, this study’s research questions are as follows:

  1. (1)

    To what extent does dissatisfaction with traditional painting tools act as a “push” factor in a designer’s intent to switch to AI painting tools?

  2. (2)

    Which specific elements contribute to dissatisfaction with traditional painting tools, enhance the attractiveness of AI painting tools, and increase switching costs?

  3. (3)

    Given these factors, how can the features and functionalities of AI painting tools be further developed and optimized to better serve the needs of designers?

Literature review

AI painting tools and switching intention

AI aims to embody the essence of intelligence and is a part of intelligent computational systems capable of human-like thoughts and actions17. Progress in AI not only contributes to economic growth but also has increasingly many uses in the field of design. AI painting tools leverage generative adversarial networks (GANs), which enable image generation from linguistic descriptions and rudimentary sketches. Examples of such tools include Midjourney, Stable Diffusion DALL-E, NightCafe, Wombo Dream, Latent Majesty Diffusion, and Google’s Imagen18. These AIGC tools autonomously generate artistic creations, serving as platforms for self-expression and reflection, thus, transcending human cognitive boundaries19. Midjourney’s creation “Théâtre D’opéra Spatial” obtained the highest honor in the digital art category at the 2022 Colorado State Fair competition18.

AI painting tools offer a multitude of features that allow designers to explore various possibilities. First, these tools serve as catalysts for inspiration and augment designers’ creative perspectives20. Second, adopting AI painting tools contributes to enhanced design efficiency. These tools accelerate the generation of design materials by efficiently analyzing large volumes of data, resulting in reduced time consumption2. Furthermore, the rapid visualization functionality inherent in AI painting tools mitigates communication barriers among multidisciplinary team members, thereby streamlining collaborative processes7. Third, AI painting tools facilitate design accessibility by providing simplified software interfaces, thereby reducing barriers to entry for both designers and the general public21. Fourth, the design outputs generated by AI are of exceptional quality22, frequently appearing identical to the work produced by humans. Finally, AI painting tools have played a pivotal role in facilitating the professional growth and evolution of designers. Collaboration with AI represents a forward-looking approach for designers seeking to innovate and adapt to changing design paradigms7. Notably, job market trends in China underscore the high demand and lucrative compensation for AI/AIGC designers.

Switching intention is a theoretical concept that denotes consumer behavior after a purchase. This refers to the willingness to alter one’s consumption behavior or transfer from a present good or service to a more desirable substitute23. Most literature on AI tools has focused on users’ adoption behavior after their initial use13,15,24, fewer studies have addressed switching intention. Switching intention is a critical topic, whenever designers cease to use painting tools they previously accepted in favor of a fresh alternative. This switch not only changes the creative process of designers but also impacts the inherent value of their design work. Given the extensive adoption of AI, it is imperative to examine designers’ intentions to switch to AI paintings. This study investigates the factors that facilitate or hinder designers’ intentions to switch to AI painting tools. It seeks to advance the technology of traditional painting tools and encourage designers to use AI painting tools.

Push–pull-mooring framework

The PPM framework, initially rooted in the study of population migration theory, was first introduced by Lee25 to elucidate migratory behavior, delineating positive factors at the destination and negative factors at the origin. However, critics argue that this PPM framework provides a broad perspective and may not sufficiently explain migration at the individual level16. Moon26 expanded this framework by introducing the mooring effect, which encompasses personal experiences or institutional influences on migration decisions. With the inclusion of mooring effects, the PPM framework presents a comprehensive model comprising push, pull, and mooring effects. Pull effects denote favorable influences that attract individuals to new locations, whereas push effects signify adverse factors that compel individuals to leave their current residences27. Individual and societal variables that facilitate or impede migration constitute mooring effects28.

The PPM framework has been widely embraced by scholars from other fields, enabling examinations of the reasons that drive consumers to migrate between products or service providers23. Recognizing its relevance, Bansal16 advocated using the PPM framework to study switching behavior and initiated research on customers’ intent to switch services. Similarly, Lai et al.29 investigated elements associated with consumers’ decisions to transition to mobile purchasing, utilizing the PPM framework as a foundational framework. Nguyen et al.23 integrated diverse medical customer-level elements to explain switching intentions in medical mobile payment applications within the PPM framework. Likewise, Cao et al.30 revealed that individuals are initially attracted to microblogging owing to its user-friendly interface, but are deterred by its limited social presence. Examining the various facets that impact consumer behavior across multiple platforms, Xu et al.31 approached their analysis from the perspective of the PPM framework.

This theoretical model was adopted for several reasons. First, the designers’ choice to switch to a specific painting tool is typically determined by their previous experience with existing painting tools, similar to the process of human migration. This inclusive understanding of migration extends to the realm of tool selection among designers, and the transition to AI painting tools mirrors the concept of migration. Second, the PPM framework is flexible and applicable to most research situations. While traditionally employed in studies of population migration, the framework transcends geographical movement to encompass various daily activities32. Finally, few studies have used the PPM framework to gather information on designers’ switching intentions to AI painting tools. Therefore, we proposed that the PPM would be a suitable framework to apply in the current study in order to fill this research gap.

Research model and hypotheses

From the PPM framework, we created a set of hypotheses to forecast AI painting tool-switching intentions (Fig. 1). The push, pull, and mooring effects were proposed to be correlated with multiple factors. The intention to switch to an AI painting tool was used as the dependent variable in the model.

Fig. 1
figure 1

Research model.

Full size image

Push effects

Traditional electronic painting tools have long served as cornerstone instruments in image processing in the design field, with software such as Adobe Photoshop and Illustrator being indispensable resources for designers. However, the emergence of AI painting tools has disrupted this conventional paradigm. The push effect is the reason why designers do not use traditional painting tools, and from practitioners’ experience, we recognize that perceived lack of technology control, inconsistent expectations, and time consumption cause dissatisfaction with traditional painting tools.

Dissatisfaction with traditional painting tools

In the literature on online learning platforms, dissatisfaction acts as a push factor, signifying negative perceptions of traditional learning modalities31. Research has indicated that dissatisfaction motivates individuals to transition to new technologies or tools33. In this study, we posit that designers contemplating switching to AI painting tools are driven by dissatisfaction with their experiences of using traditional painting tools. Thus, we proposed the following hypothesis:

H1

Dissatisfaction with traditional painting tools positively motivates designers’ intention to switch to AI painting tools.

Perceived lack of technology control

To a great extent, designers believe that they are incapable of mastering and adjusting to traditional painting tools for their design work, which is a lack of technology control34. Before using traditional painting tools, designers must learn and manipulate software commands, which frequently pose challenges in terms of control. Research suggests that perceived control influences an individual’s propensity to switch35. Conversely, a perceived lack of control over technology can induce behavioral switches. Consequently, we proposed the following hypothesis:

H2

Perceived lack of technology control over traditional painting tools positively affects designers’ dissatisfaction with such tools.

Inconsistent expectations

Based on expectation-confirmation theory, consumers’ initial beliefs about a good or service transform into assessments of its performance following usage36. Unfulfilled expectations contribute to low satisfaction with the current standards37. This study examines how perceived inconsistent expectations related to traditional painting tools contribute to increased dissatisfaction, leading to the potential intention to switch to substitute tools. Thus, a hypothesis is put forth:

H3

Inconsistent expectations regarding traditional painting tools positively affect designers’ dissatisfaction with them.

Consumption of time

The time consumption associated with traditional painting tools manifests in learning to manipulate tools and outputting design work. Proficiency in traditional painting tools necessitates an extensive amount of time and effort; however, AI painting tools provide expedited processes that allow for rapid production by using keywords as inputs. Owing to the lengthy processing time of cash payments, consumers can choose not to use this payment method38. Therefore, time-consuming processes involved in using traditional painting tools provide a strong motivation for designers to switch. Thus, the following hypothesis was proposed:

H4

Consumption of time about traditional painting tools positively affects designers’ dissatisfaction with traditional painting tools.

Mooring factors

In the PPM framework, while push and pull factors primarily drive population movement, individuals may choose to remain in their original state because of mooring factors, notably time constraints and switching costs26. Given the complexity of painting tools selection, mooring variables can facilitate or hinder the switching process. Therefore, this study identifies three mooring factors that may influence designers’ intentions to switch: habit, individual non-innovation, and switching costs.

Switching cost

Switching costs encompass various expenses, including learning, setup, financial, energy, and social risks39. Studies have highlighted the importance of switching costs in assessing switching intentions40,41,42. Consequently, when designers switch painting tools, they need to adjust to the new functionalities and face the costs associated with the transition. If the costs of switching are too high, designers may choose to use existing tools. Hence, we proposed the following hypothesis:

H5

Switching costs have a negative effect on designers’ intentions to switch to AI painting tools.

Habits

Habits refer to behaviors that are gradually developed and not easily changed38. Individual behavior that meets the description of mooring factors, such as habit, is frequently observed in studies that examine switching intentions43. Overcoming habitual behaviors requires significant effort and time, often resulting in individuals reverting to established patterns. In this context, designers may be reluctant to switch to AI painting tools because of their unfamiliarity, which ultimately reinforces their existing habits. Thus, the following hypothesis was proposed:

H6

Habits positively affect switching costs.

Individual non-innovation

Non-innovation refers to an individual’s unwillingness to try new techniques and finding it challenging to try new things44. Droogenbroeck et al.45 discovered that innovativeness is a possible element underlying behavioral intention. Innovative personal traits tend to trigger the transformation of technological constraints into adoption46. That is, a lack of innovation may amplify the perception of high switching costs, thereby preventing individuals from adopting new technologies. Therefore, the following hypothesis was proposed:

H7

Individual non-innovation positively affects switching costs.

Pull factors

According to the PPM framework, individuals will switch if alternatives offer greater value than their current option16. Notably, hedonic elements such as perceived ease of use and enjoyment influence usage intentions47. Given the efficiency of AI painting tools, this study examines perceived pleasure, efficiency, and ease of use as pull factors that drive users toward these tools.

Attractiveness of AI painting tools

In prior research using the PPM model, the attractiveness of new technologies, platforms, or tools has been mentioned as a push factor29,33,48. Chi et al.40 observed that visitors’ intentions to switch were encouraged by the attractiveness of alternatives. In this study, the designer’s preference for AI painting tools over traditional ones was based on the perceived value and satisfaction provided by AI painting tools. Thus, the following hypothesis was proposed:

H8

The attractiveness of AI painting tools positively affects designers’ intention to switch to such tools.

Perceived ease of use

Perceived ease of use, a key construct in the technology acceptance model49, indirectly enhances users’ willingness to adopt new technologies50. Owing to the apparent ease of use of new technology, people experiment with it30. Given the easy-to-use interface and the straightforward operation of AI painting tools, designers may perceive them as easier to use than traditional tools. So, the following hypothesis was proposed:

H9

Perceived ease of use of AI painting tools positively affects their attractiveness.

Perceived efficiency

Perceived efficiency refers to the positive experiential outcome individuals derive from using new technologies44. Designers may be attracted to AI painting tools because of their enhanced design capabilities and time-saving features, similar to educators who prefer smart classrooms because of their efficiency43. Thus, we proposed the following hypothesis:

H10

Perceived efficiency associated with AI painting tools positively affects their attractiveness.

Perceived pleasure

Hedonic motivation in the context of this study is defined as the intention to replace smart devices given an expected level of enjoyment or pleasure associated with owning and using a new smart device28. In this study, perceived pleasure refers to t experiencing a fun and amusement designer experience when using AI painting tools. According to Liao et al.39, users are likely to prefer learning platforms that are entertaining. The pleasure experienced when using AI painting tools may influence designers’ intentions to switch to these tools. Consequently, we proposed the following hypothesis:

H11

Perceived pleasure associated with AI painting tools positively affects their attractiveness.

Methodology

Participants and procedure

This study’s participants were Chinese designers who used traditional painting tools (e.g., Adobe Photoshop, Adobe Illustrator, and CorelDRAW) and AI painting tools (e.g., Midjourney and Stable Diffusion) to conduct design work. The sample was gathered through online questionnaires because such questionnaires enable wider sampling, being unconstrained by regional limitations. A questionnaire survey was conducted from October to December 2023. Finally, 404 designers voluntarily completed an online questionnaire. After deleting the questionnaires that took less than 60s to answer and those that reported having never used AI painting tools, 320 valid questionnaires remained. More participants were male (53.75%) than female (46.25%), and most were young (67.51%), with a bachelor’s degree (59.06%; Table 1).

Table 1 Demographic information of the participants (N = 320).
Full size table

Measurements

To guarantee sample validity, the questionnaire was gathered using a moment-positive scale with filtered questions. Simultaneously, it was divided into three parts to investigate the designers’ intentions to switch to using AI painting tools. The first part obtained basic information about the participants (i.e., gender, age, and education). The second part was designed to screen for valid samples, by asking respondents to indicate whether they used AI painting tools; a “yes” response defined a valid questionnaire in this study. The final section contained 39 questions that addressed 12 subscales. All subscales were taken from or adjusted based on previous studies.

Specifically, dissatisfaction with traditional painting tools (DIS), switching cost (SC), attractiveness of AI painting tools (ATT), and intention to switch to AI painting tools (ITS) were revised from Zhao et al.33, whose original items looked at the factors that compel online shoppers to move from live commerce to traditional e-commerce. The consumption of time (COT) and the habit (HAB) were taken from Lu and Wung38, whose original items assessed the idea of switching from cash to mobile payments. We made some adjustments, namely replacing “saving time” with “consumption of time.” The questionnaire designed by Monoarfa et al.48 was modified to include items on perceived ease of use (PEOU) and individual non-innovation (IN). The original items of the questionnaire outlined elements that encourage customers to adopt e-grocery purchases during the COVID-19 pandemic. Perceived lack of technology control (PLOTC) was measured using two questions derived from Liao et al.39. Inconsistent expectations (IE) was assessed using three questions modified from Lin and Huang37. Three questions for perceived efficiency (PE) and perceived pleasure (PP) were taken from Zhu et al.43 and Lenz et al.28, respectively.

Data analysis

The data analysis tools adopted were AMOS 26.0 and SPSS 27.0. SPSS was used to calculate the means and standard deviations and ensure that the data adhered to a normal distribution. Second, the internal consistency of each variable was assessed using Cronbach’s alpha coefficients. Composite reliability (CR) and average variance were extracted, and the measurement model was examined for validity and reliability using confirmatory factor analysis. Finally, structural equation model was used to determine whether each of the hypotheses should be supported or rejected.

Experimental results

Measurement model

Consistent with prior reports, various fit measures were calculated, including chi-square (χ2), χ2/df ratio, root-mean-square error of approximation (RMSEA), comparative fit index (CFI), Tucker-Lewis index (TLI), standardized RMR, and adjusted GFI51. These indicators were employed to evaluate the fit of the measurement model. The indices showed that the model fit the data well: χ2 = 1047.76 (> 0.5), χ2/df = 1.864 (< 5.0), TLI = 0.925 (> 0.9), CFI = 0.937 (> 0.9), RMR = 0.046 (< 0.05), and RMSEA = 0.052 (< 0.08), with every index meeting its criterion. The reliability, discriminant validity, and convergent validity of the measurement model were then assessed.

The estimated standardized factor loadings of all items are shown in Table 2; all indices exceeded the recommended criterion value of 0.50. The constructs’ CR ratings varied from 0.836 to 0.917, considerably exceeding the 0.7 criterion. The extent to which the computed variables explain the variation in the measured variables is assessed using the average variance extracted (AVE); the greater the AVE, the stronger the convergent validity of the dimension. The recommended AVE value is greater than 0.552. The model exhibited excellent discriminant validity, as shown in Table 3, in which the value on the diagonal (i.e., the square root of the AVE of each variable) is larger than all of the correlation coefficients in both the vertical and horizontal directions.

Table 2 Results of construct validity and reliability analysis.
Full size table
Table 3 Discriminate validity of the research model.
Full size table

Structural model

The results supported the hypotheses. Table 4 demonstrates that all hypotheses except H2 were supported. Inconsistent expectations and time consumption positively affected designers’ dissatisfaction with traditional painting tools (β = 0.387, p < 0.001; β = 0.207, p < 0.001, respectively), providing support for H3 and H4. Individual habits and innovativeness had a considerable positive effect on switching cost (β = 0.521, p < 0.001; β = 0.257, p < 0.001, respectively). Accordingly, both H6 and H7 were supported. Perceived ease of use, efficiency, and pleasure were associated with increased attractiveness of AI painting tools (β = 0.213, p < 0.001; β = 0.306, p < 0.001; β = 0.613, p < 0.001, respectively), providing evidence to support H9, H10, and H11. Additionally, dissatisfaction with traditional painting tools and the attractiveness of AI painting tools could exacerbate switching intentions (β = 0.139, p < 0.05; β = 0.466, p < 0.001, respectively); the higher the switching cost the weaker the switching intention (β = − 0.181, p < 0.010). The verified structural model is shown in Fig. 2. With regard to control variables, age and education have insignificant influences on switching intention, with path coefficients of 0.060 and 0.051, respectively (p > 0.05). Gender has a significant impact on the intention of AI painting tools among designers, with path coefficients of − 0.161 (p < 0.05). Specifically, compared to females, male designers have a significantly higher intention to use.

Table 4 The results hypotheses test.
Full size table
Fig. 2
figure 2

The research model with its standardized coefficients. No discernible impacts were shown by dashed lines. ∗p < .050, ∗∗p < .010, ∗∗∗p < .001.

Full size image

Discussion

The evolution of AI painting tools has markedly reshaped the skill requirements within the design field. Designers are no longer solely creators; instead, they now collaborate with technology, with their roles increasingly focused on conceptual planning and strategic decision-making. This paradigm shift not only enhances the efficiency of the design process but also raises critical discussions on issues such as design ethics and the boundaries of creativity. Moreover, advancements in AI have facilitated the development of co-creation platforms that promote interdisciplinary collaboration, enabling a more diverse group of stakeholders to participate in innovative design processes. In summary, AI painting tools have the potential to broaden both their user base and range of applications across diverse disciplines, owing to their substantial impact on design production workflows. Consequently, research and development of these tools are anticipated to expand further.

The current study’s purpose was to explore the factors that influence designers’ choice to switch from traditional painting tools to AI painting tools. Using the PPM framework, we discuss why designers accept switching to AI painting tools (e.g., Midjourney). Specifically, three factors are proposed: dissatisfaction with traditional painting tools as a push effect, including perceived lack of technology control; inconsistent expectations, consumption of time, and switching costs as a mooring effect, including habits and individual non-innovation; and attractiveness of AI painting tools as a pull effect, including perceived ease of use, perceived efficiency, and perceived pleasure. All these factors influence the switching behavior of designers.

First, regarding push factors, the results showed that inconsistent expectations regarding traditional painting tools positively affected designers’ dissatisfaction with such tools. The practical application of traditional painting tools often fails to meet the expectations of most designers, resulting in a decrease in the frequency of use of these tools. This result is consistent with Lin and Huang37. Time consumption was confirmed to positively impact designers’ dissatisfaction with traditional painting tools. Compared to AI painting tools, traditional painting tools are far slower to apply in design work. This result is consistent with Lin and Huang37. Dissatisfaction with traditional painting tools was found to motivate designers to switch to AI painting tools, which mirrors the results of Zhao et al.33.

Second, regarding mooring factors, the study shows that the habit of using traditional painting tools has a positive effect on switching costs. Persistent personal habits tend to resist change, which potentially amplifies the perceived costs associated with transitioning to new tools. Consequently, higher switching costs negatively affected the designers’ intentions to adopt AI painting tools. This finding corroborates that of Lu and Wung38 and supports Hypotheses 5 and 6. Similarly, individual non-innovation positively influenced switching costs, which aligns with Monoarfa et al.48 regarding the impact of personal innovativeness on switching behaviors in the context of e-grocery shopping.

Third, regarding pull factors, perceived efficiency, pleasure, and ease of use improved the attractiveness of AI painting tools, which is consistent with the findings of Lenz et al.28 and Zhu et al.43. In this study, the attractiveness of AI painting tools positively affected designers’ intentions to switch to such tools, consistent with Zhao et al.33.

Finally, regarding the control variables, age and education had an insignificant effect on designers’ switching intentions toward AI painting tools, which is consistent with the findings of Hsieh42. Gender negatively affected the designers’ switching intentions,male designers have significantly higher intention to use than female. In this study, Hypothesis 2 was not supported. Specifically, dissatisfaction with traditional painting tools stemming from perceived lack of technology control did not emerge as a significant factor driving designers to change their intentions to use AI painting tools. The designers did not perceive lack of control of technology as a critical disadvantage regarding painting tools. Instead, they seemed inclined to address the challenges posed by painting tools, opting to overcome difficulties rather than abandoning the use of such tools.

Implications

The study revealed that the designers’ intentions to switch were dominated by the attractiveness of AI painting tools. The primary element determining this attraction was the user’s perceived pleasure, followed by perceived ease of use. Therefore, software providers should offer useful and enjoyable services with improved usability. For example, enhancing the capacity of AI to process data and learn will expand the realm of generated content, render it more appealing, and inspire people to innovate.

According to the data, switching costs are inversely related to switching intentions for AI painting tools, and habits have a strong influence on switching costs. Consequently, one of the main factors that prevents designers from using AI painting tools is habit. Traditional painting tools have been available for more than three decades, and they have made it possible to change from drawing with a pen and paper to drawing on a computer screen; users tend to remain loyal to their chosen drawing software. To change people’s habits over time, software must be disruptively innovative and highly appealing.

These findings indicate that dissatisfaction with traditional painting tools was positively correlated with inconsistent expectations and time consumption. Furthermore, inconsistent expectations had a greater impact on dissatisfaction than did other factors. By improving systems and service quality, providers should aim to narrow the gap between users’ expectations and actual usage experience.

Limitations and future research

Despite some limitations, this study provides a sound basis for further research, as follows.

  • First, this study focused on Chinese designers who had used AI painting tools. However, because of cultural variations, the results may not be generalizable to other nations. Therefore, future research should consider the intentions of designers in various fields and with other cultural backgrounds to switch to AI painting tools.

  • Second, given the relatively short development period of AI painting tools, users’ experience with these tools remains limited, and their intention to use them may evolve over time. Therefore, future research should consider conducting longitudinal studies to investigate potential changes in users’ engagement with AI painting tools over an extended period.

  • Third, the research method was cross-sectional, in the form of a questionnaire investigation, followed by quantitative analysis of the results. Consequently, the statistical results were not comprehensive. To further explore survey results, future studies should combine quantitative and qualitative methodologies, such as interviews.

  • Forth, owing to limitations of certain objective measures, the questionnaire included a variable (i.e., perceived lack of technology control) that was assessed using only two questions, which is fewer than ideal. Additionally, the reliability test was not comprehensive. Therefore, future questionnaire designs should be more complete. Further, this study’s sample size was not large; a larger sample size would have further enhanced the result’s credibility.

Finally, while this study found that most designers adopt AI painting tools due to their perceived attractiveness, future research could focus on an in-depth examination of a specific AI painting tool to explore the precise factors contributing to this appeal. Such an investigation would provide deeper insights into emerging trends in the field of AI-generated art. Additionally, it is anticipated that the proposed model could be applied in future studies to further expand the scope and depth of research on AI art across various disciplines.

Conclusion

Despite the growing interest in AI-generated content research, studies investigating the designers’ intentions to switch to using AI painting tools are still lacking. This study adopted the model of the PPM framework on the switching intentions of painting tools in the design process. The aim was to deeply explain and predict the factors influencing designers’ switch intention to use AI painting tools. Through a questionnaire survey of 320 Chinese designers with experience in using AI painting tools, and using structural equation model to verify the direct and indirect effects between these variables. The results indicate that the attractiveness of AI painting tools (ATT) is the main factor that affects the designers’ intentions to switch, followed by switching costs (SC) and dissatisfaction with traditional painting tools (DIS). Perceived pleasantness (PP) can enhance designers’ intentions to switch to AI painting tools (ITS) through ATT. Individuals’ habits (HAB) can reduce designers’ ITS through SC. PP had the greatest effect on ITS. Furthermore, gender has a significant effect on designers’ switching intentions. Consequently, developers of painting tools may use the results of this study to improve their current products, give a user-friendly interface and designing experience, and minimize complicated instruction input. This will reduce the likelihood of designer errors and confusion, and consequently enhance designer PP. And continuous improvements in the functionality of AI painting tools will further increase the speed and accuracy of content generation. By analyzing users’ historical usage patterns, preferences, and personal styles, a more personalized experience can be provided, which in turn enhances their creative process and overall satisfaction. This research not only offers an important theoretical reference for academic studies in related fields but also provides practical guidance for developers to optimize and refine the functionality of AI painting tools in artistic contexts. Such improvements will enhance the attractiveness of these tools and increase designers’ intention to adopt them.