AI Systems Modeling Enhancer (AI-SME): Initial Investigations into a ChatGPT-enabled MBSE Modeling Assistant

Brian Johns (United States Air Force Academy)
Kristina Carroll, Casey Medina, Rae Lewark (Studio SE Ltd)
James Walliser (United States Air Force Academy)

Keywords
AI;AI-integration;ChatGPT;GPT-4;MBSE;SysML;Systems Modeling;Modeling Assistant
Abstract
As artificial intelligence (AI) becomes integral across industries, there is a growing opportunity to transform the generation of models for systems engineering. This research investigates the integration of OpenAI’s GPT-4 Turbo into CATIA Magic for Model-Based Systems Engineering (MBSE), resulting in the creation of AI Systems Modeling Enhancer (AI-SME). This study explores the comparison between models generated by AI, specifically OpenAI’s GPT, and those crafted by human systems engineers. While recognizing challenges in AI-generated models, this research underscores the potential of AI assistants to enhance the speed and accuracy of SysML model creation. Results demonstrate AI-SME’s successes in generating requirements, block definition diagrams, and internal block diagrams. Despite identified limitations such as redundancy and lack of cohesiveness in AI-generated models, the study concludes that AI-SME represents a notable advancement in AI-assisted systems engineering.