Semantic reconfiguration of AI skill taxonomies after ChatGPT: Evidence from Latin American job postings

This study analyzes whether the spread of generative artificial intelligence (GenAI) following the public launch of ChatGPT produced a structural shift in the Artificial Intelligence (AI) labor market in Latin America or whether the observed changes primarily reflect a semantic reorganization of technological skill taxonomies. The analysis is based on 14,227 job postings collected from Adzuna and Computrabajo across seven Latin American countries between February 2022 and September 2025. Using multilingual text processing and interrupted time series models, the study evaluates changes in the labeling and temporal dynamics of AI-related skills before and after the diffusion of GenAI. The results suggest that the main transformation is not a substantial increase in AI job openings, but rather a change in the terminology used to describe technological competencies. While traditional skills such as Python and Deep Learning declined in relative importance, new GenAI-related labels expanded following the launch of ChatGPT. Correlation and co-occurrence analyses further reveal the emergence of differentiated functional communities within the AI skills ecosystem, including technical-engineering, prompt-integration, and operational-infrastructure clusters. Overall, the findings suggest that the diffusion of GenAI is associated with a semantic restructuring of AI skill taxonomies rather than with a purely quantitative expansion in labor demand. The study highlights implications for curriculum design, labor-market monitoring, and skills-based hiring processes in emerging economies.