Generative AI and simulation modeling: how should you (not) use large language models like ChatGPT

Abstract

Generative Artificial Intelligence (AI) tools, such as Large Language Models (LLMs) and chatbots like ChatGPT, hold promise for advancing simulation modeling. Despite their growing prominence and associated debates, there remains a gap in comprehending the potential of generative AI in this field and a lack of guidelines for its effective deployment. This article endeavors to bridge these gaps. We discuss the applications of ChatGPT through an example of modeling COVID‐19's impact on economic growth in the United States. However, our guidelines are generic and can be applied to a broader range of generative AI tools. Our work presents a systematic approach for integrating generative AI across the simulation research continuum, from problem articulation to insight derivation and documentation, independent of the specific simulation modeling method. We emphasize while these tools offer enhancements in refining ideas and expediting processes, they should complement rather than replace critical thinking inherent to research. © 2024 The Authors. System Dynamics Review published by John Wiley & Sons Ltd on behalf of System Dynamics Society.

Integrating AI language models in qualitative research: Replicating interview data analysis with ChatGPT

“…Finally, as we recently discussed elsewhere (Akhavan & Jalali, 2023), it is important to consider that AI is not to replace the critical analytical thinking inherent to human researchers. Relying on the outcomes of these tools without understanding their limitations and without cross-checking the results can present risks to the quality of research.…”

A Visual Analytics Environment for Navigating Large Conceptual Models by Leveraging Generative Artificial Intelligence

“…In this study, we developed and used generative AI solutions in the context of modeling and simulation (M&S), which is an interdisciplinary area at the confluence of applied mathematics, computer science, and systems engineering. Generative AI can be used throughout the life cycle of M&S (Figure 1), starting with the creation of a conceptual model (i.e., a 'sketch' or 'diagram'), then moving to a mathematical or formal model (e.g., using a language for specification such as SysML), implementing the specification as a computational model via computer code (e.g., NetLogo), and performing simulations to generate insights that are shared with end-users and/or analyzed to suggest revisions in the conceptual model [5][6][7]. Our focus was on conceptual modeling, which has been the subject of several studies in generative AI either to create models from text [8][9][10] or to explain models as text [11].…”

From text to map: a system dynamics bot for constructing causal loop diagrams

“…Recent studies have shown that generative AI can be used for building simulation environments (Park et al, 2022;Park et al, 2023;Wang et al, 2023) and system dynamics models (Ghaffarzadegan et al, 2024;Williams et al, 2023). While the field of building simulation models with generative AI is expanding (Argyle et al, 2023;Dillion et al, 2023;Hamilton, 2023;Horton, 2023), the contributions of the new technology to system dynamics can go beyond simulating human behavior (Akhavan and Jalali, 2024). Our focus is on using generative AI which is in the deep-learning category for the purpose of developing CLDs.…”