Anshuman Sahoo (Indian Institute of Technology Kharagpur – Rajiv Gandhi School of Intellectual Property Law) has posted Artificial Intelligence and the Informational Infrastructure of Indian Law on SSRN. Here is the abstract:
Most legal scholarship on artificial intelligence treats AI as an object that law must regulate from the outside, asking how rules of liability, privacy or competition should be reshaped to govern a new technology. This chapter examines a different and more consequential situation: AI as a component of legal infrastructure itself. When translation systems shape what litigants say for legal purposes, when research assistants pre-frame the doctrinal field a judge surveys, and when automated systems determine welfare eligibility or evidentiary integrity, AI goes beyond merely affecting what legal institutions decide, and starts altering what those institutions are. The chapter proposes legal entropy as an analytical tool for this transformation. Legal entropy denotes the degree of uncertainty about the informational state of a legal system, measured along three dimensions: provenance, interpretation and authority. AI typically redistributes entropy across these dimensions rather than uniformly reducing or increasing it. Using three Indian case studies, judicial translation, electronic evidence under the Bharatiya Sakshya Adhiniyam, and automated administrative adjudication, the chapter argues that infrastructural AI demands a response of institutional design rather than only of external regulation. Three design principles follow from the entropy framework: provenance infrastructure, interpretive checkpoints and authority cartography.
Recommended!
To receive new posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.
Lawrence Solum