Beyond Music Score: How AI Skills and Job Fears Relate to Student Career Confidence

A survey of university music students in China found that stronger self-reported AI literacy was associated with greater career optimism, and greater AI anxiety was associated with lower optimism.

An AI tool that generates an accompaniment or suggests an arrangement can be useful to a music student developing a piece. The same technology can raise uncomfortable questions about the value of musical training and the careers that follow it. A survey of university music students in China found that stronger self-reported AI literacy was associated with greater career optimism, and greater AI anxiety was associated with lower optimism.

Published in Frontiers in Psychology, the study, "Artificial intelligence literacy, artificial intelligence anxiety, and career optimism among university music students: a cross-sectional study," by Yang Du of Xihua University examined 453 students at five undergraduate music institutions in eastern China. Perceived AI competence and anxiety each had a separate relationship with students' career expectations, even after their background and experience with AI were taken into account.

Job replacement worries stand out in the survey

Author collected anonymous online questionnaires between May and July 2026, retaining 453 of 521 submissions after quality checks. Participants were aged 18 to 23, and 67.5% were women. Their courses covered instrumental and vocal performance, music education, composition and music theory, and music technology and production, giving the survey a view across several routes into musical work.

The questionnaires measured perceived AI literacy across awareness, practical use, evaluation and ethics. They assessed anxiety about learning AI, job replacement, wider social and technological risks, and interaction with human-like AI systems. Career optimism covered expectations about professional development and confidence in reaching career goals. Average scores were 52.97 on the 15–75 literacy scale, 83.85 on the 21–147 anxiety scale and 37.87 on the 11–55 optimism scale.

Job replacement attracted the highest average anxiety rating, exceeding worries about learning to use AI, which received the lowest. Students rated their ethical understanding most highly among the literacy dimensions and their practical use skills lowest. That pattern suggests their concerns extended beyond operating unfamiliar software to questions about future opportunities and the value of their expertise.

Training and regular use reveal a clear pattern

Only 32.9% of participants had received AI training, and they reported higher literacy scores than those without training, averaging 56.66 compared with 51.16, and lower anxiety scores, at 77.58 compared with 86.93. Their career optimism scores averaged 40.25 compared with 36.70. These were differences between existing groups, so the survey cannot establish that training produced the more positive outlook.

Frequent AI users showed a similar pattern of higher literacy, lower anxiety and greater optimism. Nearly three-quarters used AI at least three times daily; those using it six or more times averaged 40.48 for career optimism, compared with 35.04 among those using it twice or less. Scores showed no clear differences by year of study or institution. Instrumental performance students reported more anxiety than composition and music theory students. Career optimism varied across specialisms overall, though no individual pair remained clearly different after correction for multiple comparisons.

Skills and anxiety each have a link to career optimism

Higher literacy was associated with lower anxiety, and every literacy dimension was positively linked with career optimism. Practical AI use and ethical understanding had almost identical relationships with optimism. Every anxiety dimension was negatively linked with optimism, with unease about interacting with human-like AI showing the strongest relationship among them. These findings suggest that technical confidence, critical judgment and emotional responses all deserve attention in discussions about students' futures.

After accounting for gender, year of study, specialism, training and frequency of AI use, literacy and anxiety retained independent links with optimism of similar strength. The full model accounted for 33% of the variation in career optimism scores, leaving substantial differences unexplained. Prior training and frequency of use no longer had clear independent associations in this analysis, making the quality and meaning of AI engagement important questions for further research.

Being comfortable with AI does not necessarily remove concerns about its effects on artistic identity or employment. Du interprets the findings through theories connecting perceived capability and technological threat with career expectations. A student can feel capable of using a tool and remain worried about what it means for a profession built on years of practice, personal expression and aesthetic judgment; the proposed psychological explanations were not directly tested.

Music education needs practical skills and honest career discussions

Author recommends teaching AI within real musical tasks, with instructors helping students evaluate generated material and retain responsibility for artistic decisions. Composition students could compare AI-generated alternatives and explain their revisions; performance students could use guided feedback activities. Music education courses could examine teaching materials and learner data privacy. Copyright, originality, cultural bias and disclosure need a place in these lessons, supported by faculty training and clear institutional rules.

Assessment could include drafts, version histories and oral explanations so teachers can see how students reached their creative decisions. Career support could give students space to discuss professional identity and the distinctive contribution of human musicians. These are recommendations drawn from the findings and wider literature, rather than teaching methods whose effectiveness was tested in this survey.

The study provides a snapshot from volunteers at five institutions in one Chinese region, limiting its reach to other settings. All main measures relied on self-report, and the literacy scale captured perceived competence rather than demonstrated ability. Counting AI use could not reveal its purpose, quality or students' dependence on it.

The survey measured expectations rather than employment outcomes, and already optimistic students may be more willing to explore AI. Research following students over time, testing educational interventions and assessing actual skills could clarify these relationships and the roles of family background, musical confidence and institutional support.