The JSW Centre for the Future of Law (JSW CFL) and the NHRC Chair on Human Rights at NLSIU co-organised a consultation on “AI, Rights and Accountability: Social Protection and Justice Delivery” with LibTech India and the Social Accountability Forum for Action and Research (SAFAR) on 23–24 September 2026.

The two-day consultation hosted discussions on the implications of technological interventions for rights, welfare delivery, public accountability and democratic participation. The consultation brought together a diverse group of participants from Karnataka, Telangana, Andhra Pradesh and Jharkhand, including representatives of grassroots and civil society organisations, accountability and research organisations, legal scholars, researchers, journalists, technologists, lawyers and members of grassroots communities. Across the sessions, participants examined how AI and digital systems are being introduced into welfare programmes, land and agricultural administration, healthcare and electoral processes, while also considering the legal and institutional frameworks required to ensure accountability.

The initial session gave an overview of how AI is currently shaping welfare, with a focus on Direct Benefit Transfer (DBT) programmes. The speakers highlighted both the potential benefits and harmful impacts of the use of AI in welfare programmes. The discussion examined how AI can be used for assessment, inspection and monitoring, as well as in the role of a watchdog or gatekeeper. The gatekeeping role of AI was particularly highlighted as raising concerns for rights-holders, as automated systems may determine access to essential welfare entitlements. Examples of apps such as OpenAgriNet demonstrated how technology could assist farmers by providing information in local languages and facilitating access to government schemes, while other interventions highlighted the risks of technological exclusion. Issues relating to Aadhaar linking, NPCI mapping, e-KYC, biometric authentication and digital attendance systems were discussed, particularly in relation to MGNREGA and other welfare programmes.

This was followed by a round-table discussion that provided important insights. Participants discussed instances where beneficiaries were unable to access welfare because of authentication failures, incorrect data or difficulties in completing technological requirements. The discussions therefore emphasised the need for effective grievance redressal, accessible information, compensation mechanisms and meaningful human intervention in decisions affecting rights and entitlements.

The subsequent session focused on the impact of AI and algorithmic decision-making on rights and public accountability, specifically in the areas of land, agriculture and health. The speakers illustrated, through examples from Telangana, how errors and omissions in digital land records could affect farmers and cultivators. The exclusion of tenant farmers from systems structured around landowners raised questions about how technological systems represent social and economic realities. There was an emphasis on developing AI systems informed by political theory and supported by effective checks and balances. Speakers emphasised that agricultural and land systems are dynamic and that algorithmic systems may reproduce errors when changing realities are reduced to fixed datasets. The round-table that followed included inputs from workers involved at the grassroots level, who provided insights into the practical challenges of implementing and relying on such technological systems.

Speakers similarly highlighted both the potential and limitations of AI in the healthcare system. Digital health records, telemedicine and AI-assisted healthcare systems may improve access and service delivery, but technological systems can also reproduce existing biases when the underlying data does not adequately represent different groups. The discussion consequently stressed the importance of diversity in datasets, privacy safeguards and accountability in the development and deployment of healthcare technologies.

The subsequent sessions provided an overview of the implications of the use of AI in welfare and its effect on democratic rights. This included discussions on the Special Intensive Revision of electoral rolls, which examined automated identification of discrepancies and the consequences of placing greater responsibility on individuals to establish their eligibility for inclusion. The speakers contrasted centralised digital systems with earlier forms of community-based verification and social audit, raising broader questions about transparency, accuracy, participation and the role of human discretion in democratic processes.

The session on legal accountability for AI and algorithmic decision-making brought these concerns together. Speakers examined the applicability of existing principles of administrative law, constitutional rights, proportionality, procedural fairness, transparency and judicial review to automated decision-making. Questions were raised about where administrative discretion resides when decisions are embedded within algorithms, databases and technical systems developed by private actors. The discussions also considered the need for explainability, accountability and meaningful participation in the regulation of AI. Participants discussed the role of AI in judicial systems and considered the opportunities presented by AI for case management, transcription, translation and legal research, while raising concerns regarding judicial data, third-party technology providers, transparency in the development of AI systems and the implications of AI-assisted decision-making. The importance of consultation with judges, lawyers, civil society and other stakeholders was emphasised.

The second day shifted towards exploring how technology, data and AI can strengthen public administration, support citizens and frontline workers, and enhance social accountability. Case studies demonstrated how technology could help identify excluded communities, improve access to government information and strengthen grievance systems. At the same time, participants stressed that technology should not replace administrative capacity or human engagement, particularly where rights and entitlements are concerned.

This was followed by a group exercise focused on strengthening accountability where AI is used. Groups examined land, agriculture, MGNREGA/VB-GRAMG and welfare, identifying non-negotiable safeguards including community consultation, transparent and dynamic datasets, human decision-making, mandatory pilot testing, accessible grievance mechanisms, non-digital alternatives and accountability for technological failures.

The event highlighted that AI governance is ultimately a question of rights, institutions and democratic accountability, requiring technological systems to remain understandable, contestable and responsive to the people they affect.