Abstract

As the research methodology, this study is the first to adopt the Hierarchical Bayesian conjoint analysis in the mismatch literature to identify and compare key qualifications and working conditions expressed by the employers and potential employees in the AI sector. The study simulates a realistic employing and applying process where an employer hires an applicant and an applicant chooses a company to work for. By doing so, the paper gathers perspectives from both employers and applicants and reveals the different perceptions between the two groups by asking them to evaluate the required job competencies and conditions from each other's viewpoints.

According to the conjoint analysis results on job competencies required in the AI sector, both employers and applicants attached importance in the following order of 'character and attitude', 'AI coding ability', 'understanding of AI theories', 'college major', and 'AI career and experience', etc. The two groups, however, diverged in perceptions regarding 'AI coding ability'. That is, companies viewed it as a far more critical attribute than an applicant did. The result points to the need for a job creation policy focused on coding ability. Also, it is necessary to collect and provide information to attract non-majors into the AI sector as high mismatch was found high among non-majors.

With regard to the working conditions in the AI sector, both companies and applicants chose 'wage (pay)' and 'growth potential of the applicant' as the two most important factors. There was, however, a wide gap between the two groups regarding the degree of the importance attached to the attributes. Moreover, the two groups had differing thoughts about the importance of the other attributes. That is, a higher mismatch was witnessed in job conditions than in job competencies in the AI sector. The result indicates that an approach oriented toward resolving mismatch in working conditions is crucial when it comes to job matching in the AI sector. Particularly notable is the willingness to pay (WTP) analysis for job conditions. The analysis quantified the values of each job condition in terms of wage(pay), thereby providing reference data when setting the size of financial support by government agencies.

Lastly, the study conducted multiple focus group interviews(FGI) with AI companies and applicants to better understand the awareness of and demand for AI sector-related job policies.