Automating AI Expert Consensus: Feasibility of Language Model-Assisted Consensus Methods for Systems Engineering

Abstract

"Expert consensus is a critical component of decision-making in systems engineering, where stakeholder input and complex trade-offs must be carefully weighed. Traditionally, consensus-building techniques such as the Delphi Method, Nominal Group Technique (NGT), and Multi-Voting have been used to aggregate expert human opinions systematically. Constant lingering challenges prove to be deterrents such as time-intensive and extensive coordination efforts required to gather Subject Matter Experts (SMEs). With the advent of Large Language Models (LLMs), there exists the potential to capture the expert knowledge and leverage AI to streamline consensus-building.

This conceptual paper explores the feasibility of LLM-assisted consensus methods in the context of systems engineering. We evaluate consensus methods based on their structure, expert interaction requirements, and compatibility with LLMs, followed by identifying which methods could be enhanced through AI-driven automation. Through a comparative analysis, we hypothesize the methods best suited for LLM augmentation or full automation and explore their potential applications in systems engineering. Finally, we discuss future research directions for both AI-driven and hybrid human-AI consensus frameworks."