Generative artificial intelligence and the transformation of the scientific research process through a critical review of the research cycle

Objective. The objective of this study was to critically examine the role of generative artificial intelligence (AI) throughout the entire scientific research lifecycle. This analysis identified the applications, benefits, limitations, and emerging tensions of the aforementioned technology. The proposed integrative perspective combined the use of the aforementioned technology with human judgment and critical validation. Design/Methodology/Approach. A narrative review with critical analysis was conducted based on an intentional selection of recent scientific literature (2020–2026) from databases such as Scopus, Web of Science, and Google Scholar. The analysis was structured according to the main phases of the research process: ideation, literature review, methodological design, data analysis, and scientific writing. A systematization matrix was employed to support the analysis and facilitate the identification of patterns, convergences, and gaps in the literature. Furthermore, generative AI tools were employed for the synthesis of documents and the organization of knowledge. Results/Discussion. The findings indicated the pervasive integration of generative AI across all phases of the research process, resulting in substantial enhancements in operational efficiency, particularly in literature reviews and scientific writing. However, limitations were also identified, including the generation of superficial content, biases derived from training data, and risks that affect methodological and analytical validity. A critical analysis of these results highlighted structural tensions between automation and human control, as well as between productivity and scientific quality. This analysis underscored the imperative for uninterrupted expert oversight. Conclusions. The application of generative AI does not result in the substitution of researchers; rather, it leads to a redefinition of their roles, which now encompass critical validation and epistemological control. Within the context of a university, the implementation of generative AI cannot be restricted; rather, it must be integrated into research training as a fundamental competence. The primary challenge confronting researchers is not technological but rather epistemological in nature. That is, the imperative lies in ascertaining that the integration of these tools does not serve to undermine but rather fortify the foundational tenets of scientific inquiry.