Synthetic Politics: A Global Typology of AI Political Actors

This paper advances the concept of synthetic politics by combining a new global dataset with a theory-driven typology of AI-based political actors. Synthetic politics refers to political processes in which artificially generated agents—such as chatbots, avatars, and generative AI systems—do not merely support human decision-makers but perform recognisable political roles, including representation, communication, mobilisation, and symbolic authority Empirically, the paper draws on an original dataset of 56 AI-based political artefacts spanning 2000–2025 The dataset reveals a pronounced post-2022 acceleration: while early cases (2000–2020) were dominated by art-led experiments and public service assistants, the period 2022–2025 accounts for nearly four-fifths of all cases and is characterised by campaign avatars, executive spokespersons, resurrected leaders, virtual dissidents, and synthetic party figureheads. Spatially, synthetic politics emerges as a multi-centric global phenomenon rather than a Western anomaly, with dense clusters in India, Japan, Denmark, China, and the UK alongside cases from Eastern Europe, the Gulf states, the Americas, and transnational movements. Building on these patterns, the paper proposes a typology organised along two analytical axes: (1) a shift from AI in bureaucracy to synthetic politics proper, where AI systems assume representative and symbolic functions; and (2) a transition from thin persona (generic assistants) to thick persona (named figures with biography, voice, and visual embodiment). We argue that persona has become a political role in its own right, redistributing visibility, responsibility, and legitimacy within hybrid human–machine arrangements. The typology thus offers a conceptual bridge between debates on algorithmic governance and emerging forms of AI-mediated representation, while providing a baseline for comparative and normative research on accountability in synthetic politics.