Generative AI influences on political attitudes and suffrage behavior: Democratic resilience as an analytical framework in the United States

Abstract

The rapid diffusion of generative artificial intelligence (AI) has introduced new challenges to democratic governance by transforming political communication, voter engagement, and information dissemination. Despite growing concerns about deepfakes, synthetic media, and AI-enabled persuasion, scholars have paid limited attention to how these technologies collectively influence voter attitudes and electoral behavior in contemporary United States elections. This study investigates the impact of generative AI technologies, including Large Language Models (LLMs), Generative Adversarial Networks (GANs), conversational bots, and synthetic media, on political persuasion and democratic resilience during the transition from the 2024 U.S. Presidential Election cycle to the 2026 Midterm Election cycle. Using a qualitative doctrinal and policy-analysis methodology supported by election-related case studies, regulatory documents, platform policies, and emerging empirical evidence on AI-generated political content, the study evaluates how AI reshapes electoral communication. The findings reveal three significant trends. First, generative AI has accelerated the shift from mass political broadcasting to hyper-personalized, data-driven persuasion that can tailor messages to individual psychological and behavioral profiles at unprecedented scale and low cost. Second, AI-generated misinformation, including deceptive robocalls, synthetic visual advertisements, and inaccurate voting-information outputs, increases information pollution and creates opportunities for voter suppression and electoral manipulation. Third, beyond direct deception, AI contributes to affective polarization and strengthens the “Liar’s Dividend,” enabling political actors to challenge the authenticity of legitimate information and thereby eroding public trust in democratic institutions. The study contributes to the emerging literature on AI governance by developing an integrated conceptual framework that links hyper-personalized persuasion, information disorder, and democratic resilience in electoral environments. The analysis demonstrates that the principal democratic risk of generative AI lies not solely in misinformation but in its capacity to undermine public trust structurally and shared epistemic foundations. To address these challenges while preserving First Amendment protections, the study proposes a multi-layered regulatory framework incorporating mandatory content provenance standards, targeted amendments to the Federal Election Campaign Act (FECA), enhanced transparency obligations for political AI content, and conditional platform accountability measures during critical election periods. The findings provide policy-relevant guidance for regulators, election authorities, technology platforms, and democratic institutions seeking to safeguard electoral integrity in the age of generative AI.