Inteligência artificial, meio ambiente e data centers: uma análise de enquadramentos nos sites Folha de S. Paulo e The Guardian

Abstract

This study aims to analyze how the environmental impacts of Artificial Intelligence are framed in journalistic narratives in Brazil and England through a comparison of the websites of Folha de S. Paulo and The Guardian. Grounded in the perspective of news framing, articulated with the Sociology of Expectations and a constructionist approach informed by a critical lens, the research is based on the assumption that journalism plays a central role in the construction of social imaginaries by selecting, emphasizing, and (in)visibilizing specific dimensions of technology. The articulation of these two theoretical traditions makes it possible to examine framing as both a cultural and a political process: cultural because it involves the condensation and circulation of socially shared meanings; political because these meanings are selectively structured by power relations and institutional interests. Methodologically, the study combines documentary research (Gil, 2008) and qualitative-quantitative content analysis (Bardin, 2011), supported by the Voyant Tools software. Through documentary research conducted on the websites of the two newspapers, 1,135 articles in Folha de S. Paulo and 602 articles in The Guardian mentioning Artificial Intelligence between September 2024 and August 2025 were initially identified. In a subsequent stage of corpus refinement, only journalistic units explicitly linking the technology to environmental impacts were selected, resulting in a final corpus of 74 articles from Folha de S. Paulo and 46 from The Guardian. The material was first subjected to quantitative analysis through the systematization of frequencies, temporal distribution, and categorization by news sections. Subsequently, a qualitative analysis guided by the framing perspective was conducted with the support of the Voyant platform, enabling the identification of lexical patterns, semantic relationships, term co-occurrences, and their contexts of use. The findings indicate that, although media coverage of the environmental impacts of Artificial Intelligence remains limited, particularly in terms of thematic centrality, the Brazilian case does not present a lower volume of coverage; on the contrary, Folha de S. Paulo published a greater number of articles than The Guardian. However, significant qualitative differences emerged. The British newspaper adopts more critical frames, emphasizing environmental risks such as energy and water consumption, whereas the Brazilian newspaper privileges approaches associated with economic development, innovation, and global competitiveness. It is concluded that these frames contribute to the construction of distinct social imaginaries and to the production of divergent expectations regarding Artificial Intelligence, thereby influencing public perceptions as well as political, economic, and environmental decision-making processes.