Automation in contract interpretation*

ABSTRACT This paper investigates whether the process of contract interpretation (also described as contract construction) can be automated. It approaches the task in two stages. The first part explains contract interpretation as a cognitive process. It demonstrates that interpretation involves: (i) the identification of arguments in favour of each interpretation; and (ii) the weighing and balancing of arguments to arrive at the construction that was probably intended. The second part of the paper explores the extent to which the process can be automated by logical design or through machine learning. It demonstrates that manual programming and data analytics can automate the interpretive process in different ways and to different extents: a machine can be manually programmed to formulate some interpretive arguments; relevant interpretive information can be identified, classified and extracted through machine learning; and data analytics can be applied to recognise argument patterns in interpretive disputes.