Large language models for air transportation: A critical review

Abstract

In the past decade, Artificial Intelligence (AI) has contributed to the improvement of various aviation aspects, including flight plan optimization, the development of autonomous systems, performing predictive analytics, as well as in passenger / crew assistance systems. The latest AI technology to potentially revolutionize air transportation are so-called Large Language Models (LLMs), which have an outstanding ability to process and generate human-like text. The application areas for LLMs cover nearly all aspects of air transportation, through language processing, content generation, and problem solving. In this study, we discuss the potential of this impact with two major contributions. First, we have performed an experimental evaluation of twelve commonly-used LLMs concerning their performance of air transportation related subjects, covering fact retrieval, complex reasoning abilities, and explanation tasks. Second, we have performed a survey among graduate students at Beihang University, a leading aviation university in China, to explore the experiences and uses of LLMs. We believe that our study makes a significant contribution towards the dissemination and application of LLMs in the air transportation domain.