Democratic Resilience Under Algorithmic Pressure: Setting a Research Agenda on Intersectional Risk Governance in Unequal Democracies

In the 2025 Bihar election campaigns, state high court courts ordered the removal of a fabricated video depicting the deceased mother of Prime Minister Narendra Modi, circulated by opposition parties to criticise the Prime Minister and influence electoral sentiment. The incident followed India’s 2024 general election, during which political parties deployed large-scale AI-driven campaigning tools, including an estimated 50 million personalised voice-clone calls, synthetic audio, and targeted digital messaging. Together, these developments have come to symbolise a new phase of democratic vulnerability under algorithmic pressure. Yet existing debates on democratic resilience and innovation focus on ex ante regulatory preparedness, while largely overlooking how these technologies interact with entrenched social inequalities in unequal democracies. This paper sets out a new research agenda for analysing democratic resilience in contested and constrained contexts, where technological governance unfolds alongside entrenched hierarchies of caste, gender, religion, language, and digital access. It argues that prevailing approaches to democratic innovation, particularly those centred on risk-based regulation, institutional design, and procedural safeguards, tend to presume relatively equal political subjects and stable access to redress. In unequal democracies, these assumptions risk obscuring how democratic harm is differentially produced and experienced. The paper proposes intersectional risk governance as a conceptual framework for future research on AI-mediated political communication. Building on regulatory governance theory, intersectionality, and vulnerability scholarship, the framework reconceptualises “risk” not as a neutral technical category, but as a form of socially distributed vulnerability shaped by institutional capacity and structural inequality. It directs attention to how regulatory and judicial responses translate collective democratic threats, such as disinformation, deepfakes, and micro-targeting, into individualised compliance failures or rights claims, often leaving systemic democratic harms unaddressed. Using India as a theoretically generative context, the paper advances an agenda for studying democratic resilience under algorithmic conditions. It argues that AI-driven political communication reshapes democratic vulnerability by intensifying asymmetries of political visibility, exposing the limits of fragmented regulatory architectures, and channeling democratic harm into individualised legal claims. In dialogue with the European Union’s anticipatory risk-governance architecture, the paper treats Europe not as a model to be replicated but as a comparative reference for assessing when democratic innovations retain coherence, and when they fracture, across unequal institutional and social contexts. By reframing democratic resilience as a distributive rather than purely institutional concern, the paper makes three agenda-setting contributions. First, it identifies inequality as a constitutive constraint on democratic innovation, rather than a background variable. Second, it highlights risk-based governance as a critical site where democratic inclusion is actively produced or denied. Third, it outlines a research programme for evaluating when regulatory responses to AI-mediated political communication strengthen democratic participation and when they risk legitimising unequal or exclusionary outcomes. The paper concludes by arguing that strengthening democratic resilience in contested contexts requires governance frameworks capable of recognising intersectional vulnerability without expanding surveillance or curtailing political participation. In doing so, it calls for a shift in how democratic innovation is studied, from institutional novelty alone toward the social distribution of democratic risk.