The cross-sectional study surveyed 425 music students across different institutions in China between March and April 2026. Researchers measured AI literacy using a validated scale covering knowledge of AI concepts, practical competence with AI tools, and understanding of AI ethics. AI anxiety was assessed with a separate instrument capturing worries about job displacement, skill obsolescence, and the growing presence of AI in creative work. Career optimism, the outcome variable, measured how confident students were about their professional prospects.

The results revealed a striking pattern. AI literacy showed a strong positive correlation with career optimism, meaning students who knew more about AI and felt more competent with it were significantly more hopeful about their futures. AI anxiety, by contrast, correlated negatively with career optimism, indicating that students who worried more about AI were less confident about their professional prospects. Both relationships held even after controlling for demographic factors.

Perhaps the most important finding concerned frequency of AI use. Although students who used AI more often showed higher career optimism in group comparisons, this relationship disappeared entirely in the regression analysis once AI literacy and AI anxiety were accounted for. The researchers concluded that how much students use AI matters less than how well they understand it and how much they worry about it. Exposure alone, the study suggests, does not build confidence.

The study also uncovered differences across musical disciplines. Performance students, who spend years developing embodied skills and personal style, showed different patterns of AI anxiety compared to students in composition, music education, or music technology. The researchers noted that disciplines centered on creative production and performance may be particularly sensitive to AI’s growing involvement in creative work, since these fields are where students form their sense of professional identity.

The researchers drew on two psychological frameworks to interpret their findings. Career construction theory suggests that confidence about professional prospects emerges from a sense of capability and agency. Social cognitive theory adds that beliefs about one’s abilities shape expectations about future outcomes. Together, these frameworks explain why AI literacy, which builds a sense of competence, and AI anxiety, which undermines it, both independently influenced career optimism among music students.

The findings carry practical implications for music education. The researchers suggest that institutions may want to combine practical AI training with attention to students’ emotional responses to AI. Simply exposing students to AI tools without building genuine understanding may not reduce anxiety or increase confidence. Instead, curricula that develop real competence with AI while addressing worries about creative work may better support students’ professional development.

The study also highlights the importance of AI ethics education. The researchers note that understanding the limitations and broader social implications of AI may help students develop a more realistic view of what AI can and cannot do in creative fields. This realistic perspective, they suggest, may support both lower anxiety and higher confidence about professional futures in music.

The researchers acknowledge limitations in their work. The cross-sectional design cannot establish causal relationships, and the sample was drawn from a single country using convenience sampling. Future longitudinal studies, they suggest, could examine how AI literacy, AI anxiety, and career optimism develop over time and whether interventions targeting these variables can improve students’ professional prospects in music and other creative fields.

The survey instrument used to measure AI literacy, the Artificial Intelligence Literacy Scale for Chinese College Students, reflects a four-dimensional conceptualization of AI competence that includes awareness, use, evaluation, and ethics. This multidimensional framing matters for interpreting the study’s findings, because it treats AI literacy not merely as technical skill but as a broader capacity to understand what AI systems do, apply them effectively, judge the quality of their outputs, and reason about their ethical and social consequences. For music students, this means that literacy encompasses questions the existing article touches on only indirectly: whether an AI-generated melody can be considered original, how copyright applies to machine-assisted composition, and where the boundaries of artistic agency lie when a model can produce convincing arrangements on demand.

The measurement of AI anxiety in the study draws on a framework that distinguishes several components, including learning anxiety, job replacement anxiety, sociotechnical blindness, and configuration anxiety. This decomposition helps explain why anxiety and literacy behaved as related but separable variables in the analysis. A student might be highly capable with AI tools yet still harbor deep concerns about what those tools mean for the value of a conservatory education, while another might feel little apprehension simply because they have never engaged seriously with the technology. The correlation between the two constructs was substantial but not so high as to collapse them into a single dimension, which is precisely what allowed the regression model to estimate their independent contributions to career optimism.

Several details of the sample and analysis deserve attention when weighing the strength of the conclusions. The 453 valid questionnaires were collected from five undergraduate music institutions in eastern China using convenience sampling, which means the participants were not randomly selected and may not represent music students nationally or internationally. The survey window ran from May to July 2026, a period when public attention to generative AI in creative industries was already well established, so participants’ responses likely reflect accumulated exposure to both enthusiastic and alarmist coverage of AI in the arts. The mean AI anxiety score of 83.85, with a standard deviation of 22.96, suggests considerable variability in how students experience technological apprehension, and the mean career optimism score of 37.87 indicates a moderately positive outlook overall rather than either widespread despair or unqualified confidence.

The demographic adjustments in the hierarchical regression included gender, year of study, music specialization, prior AI training, and frequency of AI use. That AI literacy and AI anxiety retained significant associations with career optimism after these adjustments strengthens the case that the psychological variables carry information beyond simple background characteristics or usage patterns. The final model explained 33.0 percent of the variance in career optimism, a moderate proportion that leaves substantial room for other influences, such as family expectations, labor market conditions in the music industry, mentorship relationships, and individual differences in personality or resilience, none of which were captured in the survey.

The finding that students with prior AI training reported higher literacy, higher optimism, and lower anxiety aligns with a broader pattern in the educational technology literature, in which structured exposure tends to demystify tools that might otherwise seem threatening. However, the study’s cross-sectional design means the direction of this relationship cannot be determined from the data. Students who already felt optimistic about their careers may have been more motivated to seek out AI training, rather than training producing the optimism. Longitudinal follow-up with the same students, or experimental designs that randomly assign some students to AI coursework, would be needed to disentangle these possibilities.

The theoretical framing of the study draws on Cognitive Appraisal Theory, which holds that individuals evaluate the same event differently depending on whether they perceive it as a challenge they can manage or a threat that exceeds their resources. Applied to AI, this framework suggests that two students facing identical technological change may arrive at opposite emotional conclusions based on their sense of preparedness. The observed pattern in the data, where literacy and anxiety showed correlations of similar magnitude in opposite directions with career optimism, is consistent with this account, though the design cannot confirm that appraisal processes mediate the relationships.

Context from the wider literature cited in the study underscores why music students may be a population of particular interest. Surveys of university students across multiple countries have found that a large majority now use AI for academic purposes, yet roughly half report feeling unprepared for an AI-influenced labor market. In the labor economics literature, estimates suggest that about a quarter of workers worldwide hold occupations exposed to generative AI to some degree, with the emphasis falling on task transformation rather than wholesale job replacement. For musicians, whose work involves embodiment, interpretation, and live performance, the exposure question is genuinely ambiguous: AI can generate recordings and arrangements, but it cannot yet replace the concert experience or the teacher-student relationship in a studio, and students’ career optimism may partly reflect their own judgments about which aspects of their craft remain resistant to automation.

The study’s recommendation that institutions pair practical AI instruction with critical discussion of AI’s implications for artistic identity echoes a growing consensus in music education scholarship. Prior work with vocal performance students has linked AI-related anxiety to underestimation of artistic expressiveness, and research on anthropomorphic AI features in music learning has found that students recognize educational benefits even as stronger humanlike characteristics produce psychological discomfort. These strands of evidence together suggest that emotional responses to AI in music education are not noise to be controlled for but substantive phenomena that shape how students engage with new tools and imagine their futures.

Finally, the generalizability of the findings beyond China’s music conservatories remains an open question. Cultural attitudes toward technology, the structure of professional music training, and the timing of AI adoption in creative industries all vary across regions, and the convenience sample limits confidence that the observed effect sizes would replicate elsewhere. Still, the core insight, that competence and apprehension are distinct levers that institutions can address separately, offers a practical starting point for educators anywhere who are deciding how to prepare performing artists for a profession increasingly intertwined with machine intelligence.

Subject of Research: Psychology

Article Title: Artificial intelligence literacy, artificial intelligence anxiety, and career optimism among university music students: a cross-sectional study

Article References: Du, Y. (2026). Artificial intelligence literacy, artificial intelligence anxiety, and career optimism among university music students: a cross-sectional study. Frontiers in Psychology, 17, Article 1961896. https://doi.org/10.3389/fpsyg.2026.1961896

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1961896

Keywords: artificial intelligence, music education, career psychology, higher education, AI literacy, Artificial, intelligence, literacy, anxiety, career, optimism, university