Content area
Abstract
States, tech-companies, and law-firms worldwide have been advocating to digitize and automate law to improve the course of justice and provide a fairer legal system. This dissertation problematizes this view by locating the harms of using digital technologies for legal decision-making support. A critical contribution of this research is expanding on the concept of legal pluralism – which decenters studying law from a Eurocentric and state-centric perspective – and applying it to study how law is mediated in the construction of digital technologies. I apply the concept of legal pluralism to note the failures of AI in the legal landscape by analyzing law as a broader category and including perspectives from alternate frameworks of justice such as Islamic legal thought and community-based justice systems. The goal is to show how these exclusions can result in unjust and prejudiced outputs (by humans and machines) that discriminate against marginalized communities, particularly Muslim women in South Asia and beyond.
To demonstrate my argument, I conducted two years of ethnographic fieldwork, archival research, and semi-structured interviews to examine Bangladesh’s state court and community-based non-state court called shalish to compare how law is datafied and represented in the design of two prominent technologies: 1) digital legal databases (that are used to store and retrieve data), and 2) AI Judge models (that use artificial intelligence systems to (help) generate the verdict of cases). My research demonstrates how the construction of databases and AI Judge models are not neutral, highlighting how preexisting offline social biases and modern legal epistemic frameworks of state law are reinscribed in the digitizing and automating of law; shalish, Islamic legal thinking, and rural women’s experiences of law are erased and distorted in this process. This dissertation bridges digital media studies, legal anthropology, and Bangladesh postcolonial feminist theory to show how the discriminatory outputs are tied to gendered and racialized power structures that are in part a legacy of colonialism. Dismantling the binary between online-offline spheres and studying their interdependent relationship shows how power plays a central role in the design of digital technology, and aids in reinscribing the marginalization of those who are already in the margins.
