Research report: The relationship between perceived AI replacement threat and employment anxiety among English-major undergraduates: the mediating role of career insecurity and the moderating role of career decision-making self-efficacy

Chinese undergraduates majoring in English who believe that artificial intelligence could replace their future jobs report substantially higher levels of employment anxiety, according to a new cross-sectional study of 711 students published in Frontiers in Psychology. The research, led by Suojun Zhang of Quzhou University and Hairong Zheng of Hulunbuir University, examined how a specific psychological perception—termed perceived AI replacement threat, or PAIRT—relates to the worry, tension, and unease students experience as they prepare to enter the labor market.

The study arrives against a striking demographic and technological backdrop. According to official statistics cited by the authors, China’s conventional universities graduated 12.22 million students in the class of 2025, an increase of 430,000 over the previous year, while a World Economic Forum report indicated that 40 percent of firms anticipate reduced labor demand for jobs that AI can automate and nearly two-thirds of businesses intend to hire workers with specialized AI capabilities. For English majors, whose likely career fields include translation, writing, and language services—areas with a high degree of overlap with AI application scenarios—these shifts carry particular relevance.

The researchers recruited English-major undergraduates through convenience sampling from seven universities spanning Inner Mongolia Autonomous Region, Zhejiang, Shandong, Henan, and Hebei provinces and Chongqing Municipality. The survey was administered online via the Wenjuanxing platform during March and April of 2026. Of 752 submitted questionnaires, 711 valid responses were retained after screening for incomplete answers and extreme response patterns, a retention rate of 94.5 percent. The sample was predominantly female, at 83.12 percent, and skewed toward later-stage students, with fourth-year undergraduates making up 39.38 percent of respondents.

Participants completed validated scales measuring four constructs. PAIRT was assessed with a four-item measure adapted from Brougham and Haar, including items such as “I think AI might replace the job I will do in the future.” Employment anxiety was measured with a 17-item Future Employment Anxiety Scale covering personal ability, knowledge application, career replaceability, and social relations. Career insecurity was assessed with eight items addressing both quantitative uncertainty—whether employment will continue and remain stable—and qualitative uncertainty about future opportunities, responsibilities, and job quality. Career decision-making self-efficacy, or CDMSE, was measured with the 25-item short form covering self-appraisal, occupational information, goal selection, planning, and problem solving. All scales used five-point Likert scoring, and internal consistency coefficients ranged from 0.702 to 0.959.

Before testing their hypotheses, the authors grounded the work in cognitive appraisal theory, which holds that individuals evaluate whether external events threaten personally relevant goals and resources, and that these appraisals shape emotional responses. Within this framework, PAIRT represents a threat appraisal specific to AI, career insecurity captures broader uncertainty about the security of one’s future career, and CDMSE reflects confidence in one’s coping resources—a component of secondary appraisal. The theoretical model proposed that PAIRT would relate positively to employment anxiety both directly and indirectly through career insecurity, with CDMSE weakening the first link.

The statistical analysis, conducted with SPSS 26.0, AMOS 29.0, and PROCESS 4.2, controlled for gender and grade level and used 5,000 bootstrap resamples for the mediation tests. The results supported every hypothesized path. PAIRT was significantly and positively associated with career insecurity and with employment anxiety, and career insecurity was itself positively associated with employment anxiety. In the correlation table, PAIRT correlated with employment anxiety at r = 0.551 and with career insecurity at r = 0.450, while CDMSE correlated negatively with all three focal variables.

The mediation analysis produced the study’s central quantitative finding: the estimated indirect effect of PAIRT on employment anxiety through career insecurity was 0.380, accounting for 48.29 percent of the estimated total effect of 0.787, with bootstrap confidence intervals excluding zero. In other words, nearly half of the statistical association between believing AI could replace one’s future job and feeling anxious about employment ran through a broader sense that one’s career continuity, opportunities, and quality were under threat. The direct effect of PAIRT on employment anxiety remained significant as well, indicating that the AI-specific appraisal contributed to anxiety beyond this general career uncertainty.

The moderation results added nuance. The interaction between PAIRT and CDMSE significantly predicted career insecurity, explaining an additional 0.6 percent of its variance (ΔR² = 0.006, p = 0.016). Simple-slope analyses showed the positive association between PAIRT and career insecurity weakened as CDMSE rose: the slope was 0.486 at one standard deviation below the CDMSE mean, 0.435 at the mean, and 0.385 at one standard deviation above it. The conditional indirect effect through career insecurity followed the same pattern, declining from 0.257 at low CDMSE to 0.204 at high CDMSE, and the index of moderated mediation was −0.027 with a 95 percent confidence interval of [−0.055, −0.002]. The authors caution, however, that the moderating role of CDMSE should not be overstated, since the interaction accounted for only a small share of variance and the PAIRT–career insecurity association remained significant even at higher CDMSE levels.

The team also took several steps to probe the robustness of the findings. Confirmatory factor analysis supported distinguishing the four constructs, with the hypothesized measurement model fitting better than models that combined PAIRT, employment anxiety, and career insecurity, and overall fit indices were acceptable (CFI = 0.910, TLI = 0.904, RMSEA = 0.048). Because PAIRT and the career replaceability dimension of employment anxiety share item content about potential job replacement, a sensitivity analysis re-estimated all models with that dimension excluded; the focal associations, interaction, and moderated mediation index retained their patterns. Additional checks including gender, age, grade level, and household registration as covariates, along with tests for common method bias and multicollinearity, did not alter the substantive conclusions.

The authors are explicit about the limits of interpretation. Because the design is cross-sectional and all variables were self-reported by students, the indirect and conditional indirect effects should be viewed as statistical associations rather than proof of causal or temporal links, and shared method variance cannot be completely excluded. The convenience sample of Chinese English majors may not generalize to students in other countries, institution types, or educational settings. Within those constraints, the study suggests that career support for language majors in the AI era may need to address not only technology-specific replacement fears but also the broader sense of career insecurity through which those fears appear to translate into employment anxiety, with confidence in career decision-making representing one modest individual difference worth considering in counseling and guidance programs.

Subject of Research: The association between perceived AI replacement threat and employment anxiety among Chinese English-major undergraduates, mediated by career insecurity and moderated by career decision-making self-efficacy.

Article Title: The relationship between perceived AI replacement threat and employment anxiety among English-major undergraduates: the mediating role of career insecurity and the moderating role of career decision-making self-efficacy

Article References: Zhang, S., & Zheng, H. (2026). The relationship between perceived AI replacement threat and employment anxiety among English-major undergraduates: the mediating role of career insecurity and the moderating role of career decision-making self-efficacy. Frontiers in Psychology, 17, Article 1867329. https://doi.org/10.3389/fpsyg.2026.1867329

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1867329

Keywords: artificial intelligence, employment anxiety, English majors, career insecurity, career decision-making self-efficacy, perceived AI replacement threat, cross-sectional study, Chinese undergraduates, cognitive appraisal theory, moderated mediation, translation careers, job insecurity