February 23, 2026
Impact of artificial intelligence on recruitment and selection processes
Leonardo Guimarães Coelho; Julio Cesar Ruivo Costa
Summary prepared by the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege focused on synthesis and writing.
This work analyzes the impact of artificial intelligence (AI) on recruitment and selection processes, investigating perceptions, benefits, and risks. Specifically, the research identifies trust levels in AI in selective contexts, assesses the perception of biases and risks to diversity, and examines differences in perception among candidates of different sexual orientations, contributing to the understanding of the ethical and effective implementation of technology in people management.
The rise of artificial intelligence is reconfiguring the work environment, especially people management. The application of AI in recruitment and selection (R&S) is a strategic trend that promises to optimize talent identification. Tools that automate resume screening, use chatbots, and employ predictive algorithms are increasingly prevalent in organizations seeking agility and scalability (Fernandes, 2022). This digital transformation promises to reduce operational time and costs and introduce data analysis that, in theory, makes decisions more objective.
Despite the benefits, the integration of AI in R&D raises ethical debates. The main concern, documented in the literature, is the potential for algorithms to perpetuate and amplify biases. Studies show that AI systems trained with historical data reflecting discriminatory patterns can replicate these trends, resulting in the exclusion of candidates from minority groups (Vargas and Rech, 2023). This risk imposes the need for robust algorithmic governance, with transparency, explainability, and auditing to mitigate discrimination and ensure fairness (Davenport, 2018).
The impact of AI on diversity and inclusion is a critical dimension. The promise of objective selection can be subverted if machine learning models reinforce hegemonic patterns, hindering the construction of pluralistic teams. The exclusion of underrepresented groups, even if unintentional, can consolidate barriers in the labor market, contradicting organizational inclusion efforts (Vargas and Rech, 2023). Critical reflection on the design, implementation, and monitoring of these technologies is an imperative to align innovation with social justice principles.
Candidates’ perception of the automation of selection processes is a determining factor for the acceptance of these tools. Research indicates an ambivalent view: candidates recognize AI’s potential to streamline stages, but there is resistance to the idea that technology should fully replace human judgment (Fernandes, 2022; Vargas and Rech, 2023). Human interaction and qualitative assessment are seen as irreplaceable. This perception reinforces the argument that AI’s most effective role is as a support tool, providing insights to subsidize the final decision of human resources professionals.
This study adopted a mixed methodological approach, combining bibliographic and field research. The first stage was an exploratory literature review, based on articles and academic works on AI in R&D. According to Gil (2008), bibliographic research is fundamental for building a solid theoretical framework. The selection of material sought to balance perspectives on the benefits of technology and warnings about risks associated with governance, explainability, and algorithmic biases.
The second stage was a field survey, with a structured questionnaire of 16 questions. The instrument was developed based on the discussion axes of the literature, addressing trust in AI, perception of fairness, transparency, benefits, risks, and impact on diversity. The formulation of the questions was influenced by studies investigating candidates’ perception of automation (Vargas and Rech, 2023; Albuquerque et al., 2022). The questionnaire was disseminated digitally through networks such as LinkedIn, WhatsApp, and Instagram.
The non-probabilistic sample consisted of 114 valid respondents. Participation was voluntary and anonymous. The use of questionnaires is an effective technique for obtaining standardized data, increasing the comparability and reliability of the results (Marconi and Lakatos, 2017). Sample diversity was sought, especially regarding sexual orientation, to allow for comparative analyses relevant to the research objectives.
The data analysis was performed in Excel, using pivot tables to organize and interpret the responses. This technique allowed for the tabulation of frequencies and percentages, as well as the cross-referencing of key variables, such as sexual orientation and perceptions about AI. In exploratory-descriptive studies, clarity in data organization is essential for extracting insights (Vergara, 2016). The method enabled the identification of patterns and trends, which were discussed and contextualized in light of the theoretical framework.
The profile of the 114 respondents reveals a concentration in the age group of 25 to 44 years (88.6%). This profile suggests that interaction with technologies in R&S is more intense among professionals in active career phases. The finding converges with the literature that points to the modernization of recruitment processes as a strategy to optimize time and costs, with digital tools being more adopted by generations familiar with the digital environment. Studies indicate that younger generations more easily accept AI-based solutions, as long as they are perceived as fair and transparent (Blumen and Cepellos, 2022).
The perception of justice in AI-mediated processes is a central point. When asked if AI makes processes fairer, the predominant answer (62.3%) was “It depends”. This conditionality is consistent across different sexual orientation groups, indicating a generalized distrust in the technology’s neutrality. This suggests that participants recognize AI’s potential to reduce human biases, but condition this possibility on the existence of clear rules, monitoring, and governance. This view aligns with the research by Blumen and Cepellos (2022), which associates the perception of justice with the transparency and explainability of systems.
Regarding the role of AI, the study investigated whether the technology should replace human analysis. The results were categorical: 77 respondents (67.5%) stated no, and 36 (31.6%) advocated for partial replacement. Only one participant supported total replacement. This convergence reinforces the perception of AI as a support tool, not as an autonomous substitute. The literature on algorithmic governance supports this position, arguing that the complete delegation of decisions to automated systems, without explainability and possibility of appeal, compromises the legitimacy of the process (Atanazio et al., 2021). Human supervision is crucial for the social acceptance of these technologies (Garcia and Mendes, 2025).
The impact of AI on hiring diversity revealed that the majority (56.1%) believe the outcome “Depends on how AI is programmed”. Participants do not see the technology as intrinsically good or bad, but as a reflection of its design and training data. The literature warns that historical biases in data can be encoded into algorithms, requiring audits and fairness metrics to correct deviations (Fernandes, 2022). The need for safety and ethical guidelines is not eliminated by the benefits of automation (Garcia and Mendes, 2025).
The cross-analysis between familiarity with AI and perception of its impact on diversity shows that greater technical knowledge does not translate into optimism. Individuals with advanced or intermediate knowledge also mostly expressed that the impact depends on programming, demonstrating awareness of the risks of bias. This pattern corroborates Vargas and Rech (2023), who suggest that technical expertise increases awareness of the dangers while strengthening confidence in well-calibrated solutions. The analysis by schooling showed that formal education reinforces the view of AI as a support tool, with individuals with postgraduate and higher education rejecting the total replacement of human analysis, suggesting that critical capacity requires greater transparency and supervision (Timóteo, 2024; Fernandes, 2022).
The subjective dimension of the candidate’s experience was explored. A contrast was observed in comfort when evaluated by AI: heterosexual respondents were divided (34.6% yes, 30.8% no, 34.6% maybe), while homosexual respondents showed majority discomfort (44% no, 28% yes, 28% maybe). The disparity suggests that groups historically susceptible to discrimination perceive greater risk in automated systems. Zenone and Cezar (2022) argue that trust in AI in recruitment depends on explicit guarantees of non-discrimination. Trust in AI decisions followed a similar pattern, with the majority (57.9%) expressing only partial trust, a sentiment more pronounced among non-heterosexual groups. This conditional trust requires the implementation of algorithmic governance and audits (Garcia and Mendes, 2025).
The demand for transparency was a consensus, with 98 out of 114 participants (86%) stating that companies should “always” inform about the use of AI. This requirement is fundamental for trust. Perceived risks reinforce this concern: the most cited were “lack of humanization” (80 mentions), “algorithmic biases” (78 mentions), and “lack of transparency” (50 mentions). The data confirm the analyses by Garcia and Mendes (2025) and dialogue with studies showing that the perception of risk is amplified in minority groups (Alexandre and Santos, 2023; Fernandes, 2022). The fear that algorithms reproduce structural prejudices is a barrier to the acceptance of AI (Pereira and Souza, 2021).
In contrast, the benefits of AI were recognized. The most cited were “process agility” (86 mentions), “data-driven evaluation” (67 mentions), and “cost reduction” (53 mentions). The results align with the literature that positions AI as a tool to optimize selection processes (Ferreira, 2020; Silva, 2022). Companies that adopt these technologies report efficiency gains (Cabral, 2021). However, it is crucial to note that the homosexual group presented the highest proportion of respondents who see “no benefit” (18% of the group), reinforcing that, for vulnerable groups, the risks of discrimination can overshadow operational advantages.
The analysis of the data allows us to conclude that the perception of artificial intelligence in recruitment and selection is ambivalent. The technology is recognized as an instrument of efficiency, agility, and objectivity, but this view is counterbalanced by a distrust regarding fairness, transparency, and the risk of biases. Trust in AI is predominantly partial and conditional, underscoring that human supervision is essential, especially in decision-making stages. The rejection of the total replacement of the human evaluator demonstrates that qualities such as empathy and contextual judgment are considered irreplaceable. The legitimacy of AI critically depends on the explainability of its criteria and the transparency in its use.
The breakdown by sexual orientation revealed that minority groups, such as homosexual respondents, demonstrate heightened sensitivity to the risks of algorithmic discrimination, expressing less comfort and greater skepticism. This reinforces the need to incorporate diversity as a central axis in the development of AI solutions for human resources. For organizations, the implication is clear: the adoption of AI must be accompanied by a commitment to ethical practices, algorithmic governance, and continuous audits to ensure fair and inclusive processes. It is concluded that the objective was achieved: it was demonstrated that the perception of the impact of artificial intelligence on recruitment and selection processes is profoundly ambivalent, conditioned by a strong demand for transparency, governance, and human supervision to mitigate the risks of biases, especially among minority groups.
References:
Albuquerque, E. L.; Rodrigues, F. M., Matos, G. B.; Xavier, G. S.; Hoelz, J. C. 2022. Artificial Intelligence in Recruitment and Selection: The Perspective of the Human Resources Department on its Pros and Cons. In: Management and Technology Meeting, 2022, São Paulo, SP, Brazil, Proceedings… p. 1-17.
Alexandre, D. M.; Santos, D. A. 2023. The Change in Hiring Methods Through Artificial Intelligence. Bachelor’s Thesis in Administration. Socioeconomic Center, Federal University of Santa Catarina. Florianópolis, SC, Brazil.
Atanazio, A.,;Silva, L. R. O.; Formigoni, A.; Novais, R. A. B. de. 2021. Artificial Intelligence transforming the HR of the future: a case study on technology and diversity in the labor market. Refas – Revista Fatec Zona Sul, 7(4), 1–16.
Blumen, D.; Cepellos, V. M. 2022; Brazilian School of Public and Business Administration [EBAPE]. 2022.
Cabral, J. A.. Artificial intelligence in the human resources department: a case study on AI in the recruitment and selection process. Revista Eletrônica Ânima Terra, 2021. Available at: <http://ric-cps. eastus2. cloudapp. azure. com/handle/123456789/5924> Accessed on: Sep. 27, 2025.
Davenport, T. H. Artificial intelligence for the real world. Harvard Business Review 96(1): 108-116.
Fernandes, D. L. A. Promotion of Gender Equality: The use of Artificial Intelligence in Recruitment and Selection Processes in Brazilian Companies. School of Business Administration of São Paulo, Getúlio Vargas Foundation, São Paulo, SP, Brazil.
Galvão, Y. S. M.; Câmara, R. T. T. 2023. Ethical Problems in the Use of Artificial Intelligence in Recruitment and Selection Processes and Workers’ Perception. Federal Rural University of the Semi-Arid, RN, Brazil.
Garcia, M. E. P. T.; Mendes, T. D. O. Transformations in recruitment and selection: a study on artificial intelligence in the selection process. Revista de Gestão e Secretariado, 16(2), 1-20.
Gil, A. C. 2008. Social research methods and techniques. 6th ed. Atlas, São Paulo, SP, Brazil.
Marconi, M. A.; Lakatos, E. M. 2017. Fundamentals of scientific methodology. 7th ed. Atlas, São Paulo, SP, Brazil.
Pereira, M.; Souza, A. C. Algorithmic discrimination and artificial intelligence: ethical challenges in digital selection processes. Revista Tópicos, 2021. Available at: <https://revistatopicos. com. br/artigos/uso-da-inteligência-artificial-no-processo-de-recrutamento-e-seleção-benefícios-e-desafios>. Accessed on: Sep. 27, 2025.
Silva, M. J. C. The impact of artificial intelligence on talent selection: opportunities and challenges in Brazil. Federal University of Rio de Janeiro (UFRJ), 2022. Available at: https://pantheon. ufrj. br/bitstream/11422/24539/1/MJD-Silva-min. pdf. Accessed on: Sep. 27, 2025.
Timóteo. J. F. 2024. Artificial Intelligence in Recruitment and Selection Processes: Literature Review. Bachelor’s Thesis in Administration. Institute of Applied Social Sciences, Federal University of Ouro Preto. Mariana, MG, Brazil.
Vargas, T. B.; Rech, I. The revolution in recruitment: how artificial intelligence impacts the people selection process in organizations. Bachelor’s Thesis in Business Administration. School of Business, Pontifical Catholic University of Rio Grande do Sul, RS, Brazil.
Vergara, S. C. 2016. Research projects and reports in administration. 16th ed. Atlas, São Paulo, SP, Brazil.
Zenone, A. C.; Cezar, B. B. 2020. Incorporation of Cultural Diversity in Organizations Through the Recruitment and Selection Process. In: Symposium of Research and Teaching in Administration,
Executive summary from the Final Course Work of Specialization in Digital Business from the MBA USP/Esalq
Learn more about the course; click here:
