Uneven automation: comparing authorship anxiety and generative AI use among art and design undergraduate students

Art and design education expects students to develop a sense of authorship over original work, but generative AI (GenAI) complicates how students understand and claim that authorship. This study examined whether these experiences differ across creative fields through a mixed-methods design. The survey included 2,441 GenAI-using students from seven field categories at 33 art and design institutions in China and assessed authorship anxiety, creative agency, perceived usefulness, dependence on GenAI, and frequency of use. The quantitative analysis first examined variation across all seven categories and then compared fine art with the five applied design fields. Interview data from 22 survey participants were also examined thematically. The main findings include that (1) authorship anxiety did not differ significantly across the seven categories or between fine art and pooled design; the latter comparison met the equivalence criterion at ±d = 0.20, but not at tighter bounds; (2) clearer differences emerged in how students used GenAI. Compared with fine-art students, design students considered GenAI more useful and used it more frequently; (3) across the interviews, most students described authorship as retaining control over the creative idea while delegating parts of its execution to GenAI. Specifically, fine-art students often grounded authorship in embodied making and human imperfection, whereas design students placed greater emphasis on judgement, selection, and control. We call this pattern uneven automation: concern about authorship was broadly shared, but students in different fields responded to GenAI in different ways, potentially reflecting the varying labour-market pressures surrounding their disciplines.