Data Deprivation and Constitutional Inequality

The Indian welfare state is rushing to install artificial intelligence (hereinafter referred to as AI) systems to assess who is eligible for food, employment, health and housing programmes that comprise the minimum requirements of decent human existence for hundreds of millions of people. These AI algorithms are trained and validated using digital data: biometric records, transaction histories, cell usage data, land records and formal employment records. However, the most vulnerable people in India, such as the rural poor, agricultural labourers, persons with disabilities, linguistic minorities, senior citizens, informal sector workers and members of the Scheduled Castes and the Scheduled Tribes, provide little or no such data. This study calls this condition data poverty and contends that it forms a form of structural inequality with clear constitutional ramifications. When an AI welfare system consistently excludes the data-poor, it does not simply create a technical fault; it is unconstitutional. This paper has three original arguments. First, it claims that the use of data availability as an indicator of access to welfare is a blatant arbitrariness under Article 14 of the Constitution, since the generation of data is not a rational indicator of entitlement. Second, it claims that AI-enabled exclusion from minimum welfare benefits is a violation of the guarantee under Article 21 of the right to life with dignity as defined by the Supreme Court in its foundational welfare jurisprudence. Third, and most original, it makes the case for a constitutional right to data inclusion, based on Articles 14, 15, 21 and the Directive Principles, that would oblige the State to build AI welfare programmes to actively compensate for, rather than aggravate, data poverty. The study finishes with ideas for specific legislative measures to give effect to this fundamental command