A Better Democracy or No Democracy? Imaginaries of Synthetic Democracy in AI-Driven Political Actors

This paper investigates the social imaginaries advanced by synthetic political actors and their implications for contemporary democracy. The rise of generative artificial intelligence increased the presence of synthetic political actors in the public arena. Examples include AI Steve, a system running for a seat in the UK House of Commons; Diella, the first AI minister formally sitting in a national government, in Albania; VIC, a chatbot that declared its candidacy for mayor of Cheyenne, Wyoming; among others. When examining the relationship between artificial intelligence and democracy, scholars have explored the use of AI to support civic participation, deliberation (Tessler et al., 2024) and electoral campaigns (Rubio Núñez et al., 2025). They have also offered theoretical reflections on whether AI could take on the role of politicians (Saetra, 2020; Innerarity, 2023), as well as on how willing citizens are to let algorithms make political decisions (König, 2023; Haesevoets et al., 2024). Yet despite this growing body of research, empirical studies on synthetic politics remain scarce. This paper seeks to address this gap by asking: to what extent do synthetic political actors position themselves as supportive of democracy (a “better democracy”) or in opposition to it (“no democracy”)? What democratic imaginaries do they advance, and how do these imaginaries relate to current democratic practices? The article is organized as follows. First, I present existing cases of synthetic politics through a desk-based review. I then conduct a thematic analysis of their websites, press releases, and related materials to examine the imaginaries they mobilize, drawing on social imaginaries’ theory (Castoriadis, 1997; Taylor, 2004; Jasanoff & Sang-Hyun, 2015) as well as recent empirical work on AI imaginaries (Bareis & Katzenbach, 2022). I argue that the democratic imaginary promoted by synthetic politics is not overtly anti-democratic, yet it may ultimately produce anti-democratic effects. This imaginary rests on three core assumptions: participation, technocracy, and antipolitics. First, synthetic democracy is participatory. Rather than replacing citizens, these systems frame themselves as enhancing democracy by making it more direct, participatory, and transparent. This reflects a populist echo: AI as an instrument for revealing the “authentic” will of the people. Second, synthetic democracy is technocratic. Democracy is viewed in an instrumentalist way (Norris, 2011) as a system meant to generate objective better decisions. Politics is thus not conceived as the sphere of human action (Arendt, 1956), but rather as techne, a domain of technical problem-solving (Galli, 2025). Third, synthetic democracy is antipolitical. Synthetic actors often entail a critique of the current political status quo, characterizing politicians as incapable of genuinely listening to or representing citizens. Conversely, AI is marketed as neutral, impartial, and above political divisions. In this way, synthetic actors advance a holistic view of the political community that negates the pluralist and agonistic principles on which representative democracy is built (Rosenblum, 2009).