- González Morales, Antonio
- Martín López, Milagro
- Chacón Lamela, María
The identification phase, in accordance with the PRISMA 2020 guidelines, requires specifying the databases consulted and the exact strategy used to locate studies. The Web of Science (WoS) database was selected to conduct this review. Specifically, only the WoS Core Collection (SSCI and SCI-Expanded) was targeted to guarantee that all included empirical studies had surpassed the most rigorous indexing standards of the Journal Citation Reports (JCR)—uniquely relevant to the Management and Computer Science areas addressed in this research—thereby ensuring the maximum methodological quality of the audited evidence. This delimitation was adopted to restrict the scope to a database of maximum rigor. When performing the document search, the following keyword sequence was utilized in the search engine: (“Artificial Intelligence”) AND (“Recruitment”). This combination was used as a Boolean indexing root within the Topic fields, as the granularity of sub-terms (NLP, LLM, Neural Networks) was naturally absorbed due to the high cross-indexing of WoS. These two terms were chosen as root macro-concepts precisely to prioritize the specificity of thematic matching in high-level management fields, avoiding the noise of pure engineering terms alien to HR.
Following the search, executed on January 20, 2026, a total of 438 initial records were obtained, establishing December 31, 2025, as the inclusion deadline.
However, with the purpose of obtaining a sample of recent, thematic, and fully accessible literature for complete review, the following search and selection criteria were applied directly within the platform:
- Access filter: Results were restricted to Open Access publications. This filter was applied under the principles of Open Science, given that AI research applied to human resources is heavily published in open repositories due to the speed of technological transfer and to ensure full access, auditability, and reproducibility of the sample by the scientific community, accepting any coverage limitations this might imply; the combination of the WoS Core Collection with the Open Access filter guaranteed that the final sample was auditable, reproducible, and of high normative impact, minimizing the influence of grey literature.
- Temporal filter: The analysis period was defined for articles published or indexed up to December 2025.
- Document types: Research articles were exclusively selected (Article).
- Language filter: The search was configured to include only publications in English, thus guaranteeing the linguistic homogeneity of the analysis.
- Knowledge area filter: Categories closely aligned with this study were exclusively selected (Business Economics, Engineering, Computer Science, Physiology [due to the potential use of biometrics, facial expression recognition, or neuroscience], and Social Sciences Other Topics), discarding any scientific discipline unrelated to the purpose of the research.