Artificial Intelligence applied to People Management: Potentialities and Limits of the use of Algorithms in Recruitment and Performance Appraisal

Abstract

This article analyzes the potentialities and limits of the use of artificial intelligence algorithms in recruitment and performance appraisal processes within organizations. The problem arises from the fact that decisions once assigned to human judgment now depend on statistical models trained with historical data, which reorganizes hiring, promotion and dismissal criteria without corresponding transparency about their grounds. The general objective was to examine these potentialities and limits based on Brazilian scientific literature published between 2016 and 2026. An integrative literature review was adopted, with a qualitative approach and descriptive character, searching the Capes Journal Portal, SciELO and Google Scholar through descriptors combined by boolean operators and the successive application of inclusion, exclusion, screening and eligibility criteria, resulting in twelve works examined through content analysis. The results indicate that documented gains are concentrated in operational stages, with reduced time and cost in screening applications and broader geographical reach of selection processes, while the limits appear in the automated reproduction of historical hiring patterns, the undeclared instability of evaluation criteria, the tacit requirement of technological literacy from candidates and the practical difficulty of contesting automated decisions. The discussion reveals divergence between management literature, which emphasizes efficiency, and legal and occupational health literature, which documents risks of discrimination and illness. The conclusion is that the main challenge is not technical but one of governance, since adopting such systems silently redistributes decision-making authority within organizations.