Technological Prejudice: Demonstrating the Ontological Challenge of Building a Critical Theory of Artificial Legal Intelligence

Abstract

This paper contributes to a theory of artificial legal intelligence (ALI) that harmonizes concerns for artificial intelligence (AI) bias and prejudice with 1) the critical perspective, and (2) Jacques Ellul’s critique of the “technological phenomenon”. Necessary to this contribution is an argument for the importance of ontology in understanding the multidimensionality of ALI, and critical theory’s ability to deal with this multidimensionality. First, the paper introduces critical theory and some of its tenets. My focus then is critical legal studies (CLS) and their contentious relationship with the ontological issue of instrumentality. I emphasize that one way a theory of ALI can engage with this critical theme is through an ontological classification of AI. I propose two classifications: AI as a tool and AI as an ideological phenomenon. Each classification is attributive of a certain autonomy to AI and telling about a potentiality for domination a critical theory of ALI should recognize, deconstruct, and challenge. Ellul’s argument that the technological phenomenon is “autonomous” informs this part of my argument. I then discuss the concept of “prejudice” and find that, considering the ontological classifications, prejudice is visible in more than one form. Although the “algorithmic bias” approach is adequate for AI as a tool, it does not account effectively for another form of prejudice rooted in technology. I call it technological prejudice.