The Role of Language Prediction Models in Contractual Interpretation

Abstract

Boilerplate language tends to plague most parties’ understanding of contractual clauses. Keeping this semantic difficulty in mind, the OpenAI Foundation created Generative Pre-trained Transformer-3 (GPT-3) in order to, inter alia , simplify and personalize contracts. Essentially, GPT-3 is fed with copious bytes of information from publicly available sources to assist users in language structure and simplification. Arbel and Becher believe that “smart readers”, powered by GPT-3, can be used in mobile and computer applications to effectively read contracts and explain their content. They have shown that GPT-3 can provide succinct and personalized summaries of contractual clauses, thereby demonstrating its potential to revolutionize the understanding of contracts. This chapter will first explore this technology and its inevitable impact on the future of the legal profession, including the possibility of courts imposing a heavier burden on parties based on the underlying assumption that there is a “duty to read”. Thereafter, it will delve into the uncertainties involved with such technology, essentially relying on its impact on the readability of standard form contracts. The chapter intends to contribute to the limited literature on the impact of language prediction models on contractual interpretation.