Modeling public opinion formation processes using multi-agent AI systems: practical experience
Year: 2026
Type: article
Abstract: The relevance of the research is due to the rapid strengthening of artificial intelligence as personal assistants to users, as well as a tool for analyzing and predicting social processes. The purpose of the article is to determine the degree of applicability of multi-agent swarm systems using the example of MiroFish to the construction of the public opinion formation process. The empirical basis for writing the work was a series of 4 author's simulations of public discussion in various scenarios: the introduction of an age limit on the use of AI, the closure of the plant, the reaction to «trash bloggers» and spontaneous online discussion. The samples are synthetic in nature, their size ranges from 100 to 2000 agents (total n=2400). It has been established that the MiroFish system, configured to work with LLM YandexGPT, can simulate the process of constructing public discussions. But this possibility is only partial, since the system's capabilities are limited by the technical features of LLM. The novelty of the work lies in the empirical verification of the sociological validity of the use of such services, the practical significance lies in the systematization of the possibilities and limitations of using the method as an additional tool for sociological research.
Author: M. L. Gorshenin
Institution: Kazan Federal University
Language: English
Subfield: Statistical and Nonlinear Physics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7215922485