From Causal Inferences to Predictive Analytics: Using AI to Settle on Damages

Abstract

From the perspective of both plaintiffs and defendants, the measurement of damages quantum is of the utmost importance. It is surprising to see this process left entirely to the court’s discretion, as quantum is traditionally considered a question of fact. Quantifying damages presents significant challenges due to the subjective nature of court discretion, leading to uncertainty for both plaintiffs and defendants. This research addresses this issue by examining difficult-to-quantify contract damages through empirical and comparative methodologies. Based on prior studies on French civil law and American common law, this empirical study involved quantitative analysis of various contract cases. Methodological advancements, including Machine Learning (ML), and Natural Language Processing (NLP) techniques, facilitated automated extraction and analysis of key variables. With a focus on overcoming sample size limitations and enhancing accuracy, this study achieved a classification accuracy of over 85% for identified essential variables. The more recent integration of generative AI and Large Language Modeling marked significant progress in quantifying damages. I conclude with recommendations for sustainable management practices in this field.