Practical implementation of numerical analysis automation in geotechnical engineering with generative AI assistance: advantages and challenges

In recent years, numerical analysis has emerged as a powerful tool for tackling a wide range of geotechnical problems, including complex ground-structure interactions. Advances in software and hardware have expanded the capabilities of numerical packages, enabling diverse design options and comprehensive output reporting. Despite these advancements, the process of developing numerical models and extracting results remains resource-intensive and time-consuming, often leading to increased costs and heightened risk of human error. Automating these procedures can accelerate the design process, reduce costs, and allow expert resources to focus on more technical tasks. Numerical packages often include built-in programming languages, such as Python, which can be used to develop automation codes. However, geotechnical practitioners frequently encounter challenges such as time constraints and a lack of programming skills. Furthermore, companies are generally hesitant to hire experienced programmers solely for developing automation codes. This raises the question of how geotechnical engineers can acquire the necessary skills to efficiently develop automation codes for their routine design tasks. The recent advancements in Artificial Intelligence (AI), particularly Generative AI (Gen AI), offer promising solutions for learning and coding. This paper outlines a methodology employed by geotechnical practitioners with limited coding experience to create practical automation codes. It highlights how Gen AI platforms, such as ChatGPT, act as effective virtual assistants, accelerating the coding process and enhancing productivity. The paper presents three case studies demonstrating the advantages and benefits of automating numerical procedures. Additionally, it discusses the challenges associated with using Gen AI platforms as virtual assistants for geotechnical practitioners.