Artificial Intelligence Powered Integrated Recruitment Support System for Migrant Workers: Job Matching, Fake Agency Detection, Skill Mismatch Prediction and Conversational Guidance Automation

Abstract

The study proposes a novel suggestion of a recruitment support system based on artificial intelligence (AI) to augment the security, equity, and efficacy of international job recruitment of foreign employees. The migrant workers are likely to have dubious job information, corrupt recruiting firms, lack of expertise in the activities, and language barriers that increase the risk of exploitation and poor working outcomes. The proposed system is based on the four main components, which are: employment congruence, prediction inability of the skills, detecting the fraudulent agency, and multilingual chat-support as the tool of increasing the transparency and informed selection. It was tested on 2000 profile of workers, 1500 overseas jobs and 300 recruitment agents with 80/20 split, which is 2000 workers profiles, 1,500 foreign job openings and 300 recruitment agents. The accuracy of the job matching prediction by the use of the TF-IDF-cosine similarity (Top-1) was 0.86 and the mean absolute error of the skill mismatch prediction using the regression analysis was 0.41. It was trained on machine learning classifiers, which classified fraudulent recruitment agencies with F1 score of 0.89 being discovered utilizing confirmed agency records and complaints. Regardless of the limitations linked to the quality of data, incomplete worker profiles, and language barriers, the results suggest that AI-based hiring systems have great possibilities of reducing the possibility of exploitation, better access to reliable information, and safer and fair overseas hiring.