Algorithmic Governance and New-Generation Policing in India: The Changing Dynamics of Discretion, Autonomy and Control

Start: May 2024

The regulation of social life through artificial intelligence is fast becoming a global phenomenon. From predictive policing to revenue administration, algorithmic governance is reshaping the everyday practices of states around the world, promoting epistemologies that treat the world as measurable, controllable, and best managed through data-driven decision-making rather than human judgement. While proponents argue that this brings efficiency and objectivity to governance, critical scholars have documented how such systems can encode and amplify existing inequalities of race, caste, class and gender, disproportionately affecting historically marginalised communities. 

India is no exception to these transformations. The current wave of AI-led state reform however must be understood against the backdrop of the liberalisation of the 1990s and the technocratic imaginaries that have long shaped governance in India; from colonial administration to post-colonial development. Algorithmic governance, with its well-documented ‘black box’ problem, represents a new form of exclusive expertise that stakes authoritative claim over effective governance, displacing local knowledge and contextual judgement in favour of ‘objective’ data. 

This research project analyses algorithmic governance in India by studying algorithmic transformations in policing in the state of Kerala. The research focuses on understanding policing and democratic state practices in Kerala with particular focus on the changing dynamics of discretion, autonomy and control among state personnel.  

The study draws on three interconnected sources: dominant discourses on AI governance and predictive policing across research, policy, and public debate; state laws, policies and orders that codify the regulation of algorithmic systems; and on-the-ground reportage and interviews with police personnel, local state actors, and the technology companies/institutions building tools for algorithmic policing. Kerala offers a particularly instructive case, as the state police have chosen to develop in-house software rather than rely on commercial tools. This seemingly minor difference reveals how state institutions actively negotiate their own role in society, even as they adopt the data-driven frameworks that are redefining governance worldwide. 

By tracing the relationship between perception, codification and everyday practice, this project seeks to unearth the operational power relations that co-produce algorithmic governance, and to understand what is at stake for democratic governance and human rights when discretion is reframed as bias, and data as truth. 

This research project contributes to CGHR’s research theme on Global Experiences of Algorithmic Governance, bringing perspectives from a postcolonial state in the Global South in conversation with dominant Global North perspectives on AI governance. In doing so, it advances critical knowledge on how algorithmic systems are experienced and contested in diverse global contexts, with implications for democratic accountability, predictive policing, and human rights.