Between AI Literacy and Public Participation: Perspectives from AI Experts in Brazil

Despite artificial intelligence (AI) being increasingly integrated into everyday life, studies indicate that the general public remains relatively unfamiliar with AI, often lacking a clear understanding of how this technology works or how it is used in daily contexts (Carvalho et al., 2025; Nader et al., 2022). This is particularly relevant in debates on AI governance and regulation, where public participation is frequently presented as an essential element of democratic legitimacy (Ulnicane et al., 2020). The literature on AI governance identifies several challenges. It highlights difficulties in practical applicability resulting from limited technical expertise within government bodies (Kyriakou & Otterbacher, 2023) and from the formulation of regulatory strategies with restricted involvement of the technical community (Khan et al., 2023). In addition, it suggests a gap in understanding how concrete governance practices incorporate public perspectives and multi-stakeholder approaches (Machado et al., 2023). Another critical issue is that these debates have been conducted predominantly from Global North contexts, with limited attention to the institutional, social, and economic realities of the Global South (Kim et al., 2023; Ulnicane et al., 2020). The challenges of AI governance are amplified in the Global South, where historical power asymmetries persist, and technological infrastructure remains limited (Instituto Igarapé, 2024). Deeper socio-economic inequalities make protection against algorithmic bias and digital exclusion more urgent, as the misuse of technology in education, health, and public safety could further exacerbate existing vulnerabilities. In this context, Brazil serves as a notable case study. Despite ranking 27th globally in startup ecosystems (StartupBlink, 2025) and leading AI development in Latin America (Soares Seto, 2025), the country lacks dedicated federal AI legislation. Studies also indicate low levels of political trust (Edelman, 2025; OECD, 2024) and scepticism among technology professionals regarding the legitimacy and feasibility of AI regulation in the national context (Soares Seto, 2025). Against this backdrop, this paper investigates how AI experts in Brazil perceive the general public’s AI literacy and its role in AI governance and regulation, addressing three sub-questions: (SQ1) how these experts describe public perceptions of AI; (SQ2) what level of AI literacy they identify among the general public; and (SQ3) what views they express regarding public participation in AI regulation. The study adopts a qualitative approach based on semi-structured interviews with AI experts in Brazil, that is, actors with advanced knowledge of AI who play an active role in one or more stages of the AI system lifecycle. Finally, this paper outlines a central provocation: although public participation is widely regarded as essential to the democratic governance of AI, being associated with democratic legitimacy, equity, and ethical alignment between technology and societal expectations (Ulnicane et al., 2020; Machado et al., 2023), it risks becoming merely symbolic and performative when not accompanied by adequate levels of AI literacy. By focusing on AI literacy as a dimension of digital competencies, the study engages with debates on the second-level digital divide and its implications for public engagement in contexts of digital governance, offering empirical evidence from a Global South context.