Artificial intelligence, scientific inequality and effects on knowledge production: how AI widens asymmetries in research

Abstract

Unlike other digital technologies, artificial intelligence (AI) has been incorporated into the field of scientific research under a predominantly optimistic narrative. Within this framework, AI is presented in both formal and informal discussions as a technology capable of democratizing research processes, from research design to data analysis, reducing barriers to entry and benefiting traditionally disadvantaged researchers or ecosystems. This essay challenges this premise, arguing that AI does not necessarily reduce inequalities in the production of scientific knowledge, but rather tends to widen them. The central argument is that its effective use depends on pre-existing capital and skills that are unevenly distributed among researchers and institutions. Consequently, the expansion of AI introduces a new layer of value that reinforces cumulative advantages and accentuates the gap between those who know how to integrate it effectively into their research processes and those who do not. Using the specific case of communication research, this essay explores the ambivalent effects of this technology and argues for the need to strengthen methodological and theoretical training as a necessary condition to prevent AI from consolidating a false sense of security regarding competence and a standardization of knowledge production that leads to increased productivity detached from quality.