Legal Positivism & AI


Associated Researcher


By exploring the relation between legal theory and machine learning, this project contributes to providing a basis for further discussions on the transformation of the legal system based on the ongoing digitization.

The ongoing digitization of society and the connected discourses and imaginaries are not just constituting a general phenomenon, but also impact specific arenas of social life. The law, as a central institution in the constitution of states and social systems, already is and will also further experience transformations by new digital and data driven technologies and Artificial Intelligence.

The project “Legal Positivism & AI” is concerned with these relations between new legal technologies (e.g. case risk assessments) and the law as an established social institution. Imaginaries of neutral algorithms that realize the law in a standardized way thereby are strangely at odds with an ongoing debate within legal theory, namely between legal realists and legal positivists. At the same time interdisciplinary studies showed that algorithms themselves are reliant on manifold practices of ‘making the algorithm’ that include normativity, questioning the narrative of its objective and neutral nature. This raises questions, what theoretical idea of the law is actually realized when algorithmic applications are increasingly being used within the legal system. This question is further being complicated by the fact that neither the legal realism nor the legal positivism exists. Deconstructing the socio-technical epistemology of AI creates an interesting tension between the imaginary of data-driven technologies, the actual and practical realization of AI applications, and their relation to different versions of legal positivism and realism. Drawing from STS, Sociology, and Legal Theory, the project therefore reconstructs the different discussions within legal theory, comparing central epistemological and ontological positions and relates them to the narrated and actual epistemology of AI.