The Competence Paradox: Negotiating Ease, Risk, and Creative Identity in Text-to-Image Generative AI Use Among Art and Design Students

Synthesis: Liu, Meng, and Zhang (2026) examined technology acceptance of text-to-image (T2I) generative AI in art and design education from both educators' and students' perspectives, using a modified exploratory sequential mixed-methods design (QUAL-QUAN-qual). Based on instructor focus groups, a survey of 417 college students, and semi-structured interviews, they found that performance expectancy, social influence, novelty value, and creative competence positively influence behavioral intention, while the negative effects of effort expectancy and facilitating conditions reflect students' shortcut-oriented use in coursework. They identify a competence paradox: although creative competence supports behavioral intention, it may also lead to more selective or restrained actual use as students negotiate authorship, originality, and skill preservation.

Key Findings

• Performance expectancy, social influence, novelty value, and creative competence positively influence behavioral intention to use T2I generative AI in art and design coursework.
• The negative effects of effort expectancy and facilitating conditions are interpreted in light of students' shortcut-oriented use of T2I tools in coursework — ease and availability enable quick output generation rather than sustained engagement.
• Students with different levels of competence perceive distinct risks across task stages, helping explain the lack of significant translation from intention and creative competence into actual use behavior.
• A competence paradox emerges: creative competence positively supports behavioral intention but may also lead to more selective or restrained engagement in actual use, as students weigh authorship, originality, and skill preservation against efficiency.
• The study reconceptualizes T2I adoption as a dynamic negotiation between diverse student profiles and technological evolution, extending technology acceptance models to creative education where utilitarian predictors alone are insufficient.

Study Design & Method

The study used a three-phase sequential mixed-methods design (QUAL-QUAN-qual). In phase one, instructor focus groups identified key constructs and informed a contextualized technology acceptance framework. Phase two administered a questionnaire survey to 417 college students. Phase three used semi-structured interviews to explain unexpected quantitative results. Analyses examined how motivational and contextual factors (performance expectancy, effort expectancy, social influence, facilitating conditions, novelty value, creative competence, risk perceptions) translate into behavioral intention and actual T2I use, with attention to how students' interpretations of risk and developing creative identity shape adoption.

What this means for practice

• Instructors. Assess the studio process — sketches, drafts, prompt logs, and revisions — alongside the finished artifact, so T2I use is judged as part of an iterative creative workflow rather than an efficient substitute for it.
• Instructors. Teach prompt crafting and output evaluation as explicit studio skills, giving students criteria to judge a generated image and to say why it fails before they accept it.
• Instructors. Design assignments that require revision across multiple stages, because ease and immediate visual feedback are what convert novelty into a completion shortcut.
• Learners. Keep an explicit record of what you delegate to the tool and what you keep for yourself, so decisions about authorship, originality, and skill preservation stay deliberate rather than habitual.

Limitations

The study draws on a single institution's art and design students and faculty, bounding generalizability to other creative disciplines and contexts. The mixed-methods design, while rich, relies on self-report for acceptance and intention constructs, and the explanatory qualitative phase is limited to explaining quantitative results rather than independently establishing mechanisms. The dynamic, evolving nature of T2I technology means findings may not fully generalize to future tool capabilities.

Citation

Liu, Y., Meng, M., & Zhang, Y. (2026). The competence paradox: Negotiating ease, risk, and creative identity in text-to-image generative AI use among art and design students.