Skill mismatch: exploring the impact of generative AI in detecting biases and discrimination in job advertisements

The Human Resources field, on a national and global scale, faces increasing challenges related to Skill Mismatch, a phenomenon in which candidates’ profiles do not align with labour market demand. The contributing factor analysed in this paper is the lack of inclusivity in job advertisements, as non-inclusive language and biases in the selection criteria may discourage applications from underrepresented communities. Ensuring compliance with Diversity & Inclusion (D&I) principles in job postings is essential to fostering fair hiring practices and promoting employment equity. While AI-based technologies are finding increasing applications in the recruitment process, thus becoming part of the research topics in HR management, most existing tools primarily focus on gender discrimination, underestimating other critical aspects of D&I. This study explores how an integrated approach, based on a broader range of aspects of D&I European principles, may enrich AI-based tools to identify, but also solve, biases in job advertisements, enhancing inclusivity and employment equity for both candidates and recruiters. The approach to the problem has been broken down into two main phases. First, guidelines were developed based on European non-discrimination regulations, providing a structured framework for evaluating job advertisements. Then, a Generative AI-based software was designed to screen job postings for inclusivity and compliance. The findings underscore the potential advantages of AI-driven automation in screening job advertisements, reducing human biases, resource commitment, and processing time while ensuring adherence to D&I principles.