Deepfakes e a crise de confiança na democracia: cenário sobre a influência da desinformação por IA nas Eleições de 2026

Abstract

The 2026 elections will take place in a scenario of technological maturity of deepfakes – synthetic videos and audios generated by artificial intelligence with a high capacity to deceive. This phenomenon brings back into debate the fragility of public trust in electoral contexts. This undergraduate thesis aims to analyze possible scenarios regarding the influence of deepfakes on the Brazilian electoral process, considering different institutional responses. The research problem is: what scenarios can unfold for the 2026 elections based on the influence of deepfakes on voter trust and public debate? The research adopts a qualitative approach, based on documentary analysis of reports from the Superior Electoral Court, the Office of the Comptroller General, and international organizations, in addition to examining real cases of deepfakes identified between 2024 and 2025. Based on the literature on disinformation and cognitive biases, it discusses how these contents affect voters' perception. Finally, prospective scenarios are constructed that point out possible consequences of the action (or omission) of public management in confronting artificial disinformation. It is concluded that the future of electoral integrity will be directly related to the institutional capacity to anticipate these new forms of manipulation.