AI CLAIM CORRESPONDENT

Abstract

This research describes the creation and implementation of an improved artificial intelligence (AI) model utilizing ChatGPT 3.5 Turbo that is especially designed to better understand and apply FIDIC clauses in the construction industry. Due to ChatGPT 3.5 and ChatGPT 4.0's inadequate understanding of these intricate legal texts, OpenAI's Text-Embedding-Ada-002 and custom Python code were used to enhance the ChatGPT 3.5 Turbo model. The study developed gradually; first, our aims and research questions were defined by a thorough examination of the literature. In order to ensure legal significance, certain aspects had to be perfected and embedded throughout the second phase of practical development. Interaction with this AI model was made possible through the development of a web application that focused on scenario-based testing and was based on the 1987 and updated 2017 FIDIC provisions. The entire public was also given access to a personalized GPT. This step-by-step process refined the AI's capabilities and ensured that it met modern requirements; the result was the integration of the AI into a commercial online platform along with a specially designed GPT for real-world applications. The methodology, the development process, and results of integrating artificial intelligence (AI) with legal document interpretation in the construction sector are all covered in this research.