Application of Natural Language Processing to Automatically Identify Exculpatory Clauses in Construction Contracts

Abstract

Construction projects, by their very nature, are capital-intensive and risk-prone. Risk allocation and its management are therefore, integral parts of construction operations. A reasonably drafted contract allocates risk to the party that is best capable of handling it. Imbalance in risk allocation, however, is a common contract feature. By including exculpatory clauses, risks are often transferred without assessing a party’s ability to handle them. This bias in risk allocation promotes an adversarial relationship. Therefore, it is essential to identify such clauses before signing a contract. While it is possible to identify them by reading the bid documents, the manual process is time-consuming and often not pursued considering the lack of time in the bidding stage. Automation of such tasks can therefore aid a manager’s decision making. However, it requires tools that can quickly and reliably identify and extract exculpatory clauses. This study developed a natural language processing (NLP) model as a proof of concept to identify exculpatory clauses automatically. The developed model demonstrates that NLP is a potentially useful tool, aiding decision makers in refining their negotiation strategies before signing a contract.