Task fit and enjoyment tied to designers’ AI image tool adoption

Generative artificial intelligence is reshaping how designers generate concepts, explore visuals, and iterate on solutions, but why individual designers actually intend to use AI-generated image tools has been difficult to explain through efficiency or ease of use alone. A study published in BMC Psychology argues that the answer lies partly in how well the technology matches the specific demands of design work. Combining two established behavioral frameworks, the unified theory of acceptance and use of technology (UTAUT2) and the task–technology fit model, the researchers analyzed questionnaire responses from 338 designers who reported prior use of AIGC image-generation tools, adding personal innovativeness and design ethics to the standard model.

The team used two complementary statistical approaches. Partial least squares structural equation modeling (PLS-SEM) tested the average strength of associations between variables across the whole sample, while fuzzy-set qualitative comparative analysis (fsQCA) searched for combinations of conditions that appear together in cases of high or non-high behavioral intention. This dual design matters because an average association can hide very different routes to the same outcome: two designers may both intend to adopt an AIGC tool, but for entirely different constellations of reasons.

The PLS-SEM results point to task–technology fit as a central variable. Fit between what the tool can do and what the design task requires was significantly and positively associated not only with behavioral intention itself, but also with performance expectancy, effort expectancy, and price value. In turn, performance expectancy, effort expectancy, hedonic motivation, price value, and habit each showed significant positive associations with behavioral intention. Notably, the direct paths from social influence, facilitating conditions, personal innovativeness, and design ethics did not reach statistical significance. In this study, design ethics was defined mainly as awareness of copyright, intellectual property, creative attribution, and transparent use.

The configurational analysis told a more nuanced story. The baseline fsQCA identified three configurations associated with high behavioral intention. Two of them, labeled A2 and A3, combine hedonic motivation with task–technology fit, facilitating conditions, habit, and other conditions, and they remained under a more stringent case-frequency threshold. The third, A1, contained six cases and showed relatively low raw coverage; when the researchers raised the case-frequency threshold, this configuration disappeared, supporting an exploratory interpretation rather than a robust one. In other words, the enjoyment designers get from these tools appears in combination with task fit, supportive resources, and established habits rather than on its own.

Just as revealing is what the analysis found for designers who did not show high intention. Configurations associated with non-high behavioral intention were not simple reversals of the high-intention pathways. The absence of a condition in one group does not automatically mean its presence produces adoption in the other, a finding that underscores why the researchers combined an average-effects model with a case-based method in the first place.

Overall, the study suggests that designers’ evaluations of AIGC image-generation tools are grounded in the fit between technological capabilities and the actual demands of design tasks, while hedonic motivation, task–technology fit, facilitating conditions, price value, and habit occupy different positions across the identified configurations. The authors propose that these findings may serve as reference points for AIGC tool development, for managing design organizations, and for cultivating designer capabilities. The evidence base carries clear limitations: the data are cross-sectional and self-reported, the sample consists of designers with prior tool experience, and the reported associations describe relationships rather than proven causal effects. The fsQCA results are also sensitive to how case-frequency thresholds are set, as the fate of configuration A1 demonstrates. Even so, the combined picture offers a concrete starting point for understanding which conditions travel together when designers decide whether generative image tools belong in their workflow.

Subject of Research: Factors associated with designers' behavioral intention to adopt AIGC image-generation tools, analyzed with PLS-SEM and fsQCA

Article Title: Explaining designers’ adoption of AIGC image generation tools: A configurational analysis integrating PLS-SEM and fsQCA

Article References: Zhang, J., Yang, R., Song, H., & Jiang, Q. (2026). Explaining designers’ adoption of AIGC image generation tools: A configurational analysis integrating PLS-SEM and fsQCA. BMC Psychology. https://doi.org/10.1186/s40359-026-05675-z

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05675-z

Keywords: AIGC, image generation tools, designers, technology adoption, UTAUT2, task-technology fit, PLS-SEM, fsQCA, behavioral intention, design ethics, hedonic motivation, BMC Psychology