Our goal is to initiate a conversation around a foundational divide in artificial intelligence: the relationship between rule-based (deterministic) systems and the probabilistic (pattern-driven) systems underlying most machine learning. This dichotomy is often assumed – and even bridged through neuro-symbolic strategies, but it is only rarely explicitly interrogated. We believe that this divide is not just a matter of current IT strategy in legaltech and other applications; it shapes the structure of many decision-making systems, including traditional jurisprudence and human cognition.

Our discussion will range from exploring the theoretical basis for using the different strategies as decision making tools, through practical considerations such as development and implementation costs and impacts, and on to examples of specific applications that choose and blend the core strategies.  The session will be guided as it progresses through these topics but will not have formal panels or speakers. This conversation is intended to be a starting point, allowing experts in different pieces of this inquiry to share knowledge and find connections that can help advance our understanding of these critical foundations of contemporary life. Through this exchange, we hope to spark collaborations, research initiatives, and projects that deepen our understanding of how these systems shape modern decision-making.

Organizers:
Jeannette Eicks, Associate Dean and Professor Law, The Colleges of Law
Oliver Goodenough, Stanford CodeX Affiliate, Research Professor, Vermont Law and Graduate School and Senior Lecturer, Thayer School of Engineering at Dartmouth College