Simulation of Language Evolution under Regulated Social Media Platforms: A Synergistic Approach of Large Language Models and Genetic Algorithms

Jinyu Cai, Yusei Ishimizu, Mingyue Zhang, Munan Li, Jialong Li, Kenji Tei · 2025

Read Simulation of Language Evolution under Regulated Social Media Platforms: A Synergistic Approach of Large Language Models and Genetic Algorithms by Jinyu Cai, Yusei Ishimizu, Mingyue Zhang, Munan Li, Jialong Li, Kenji Tei (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-19193.txt.

Description

Jinyu Cai and five co-authors simulated how slang evolves to dodge social media rules, in a 2025 paper. Platforms restrict some content, and users invent coded ways to keep talking. Their framework uses language model agents in two roles. User agents adapt their wording, while supervisor agents judge rule breaches.

Language plans were split into a constraint part and an expression part. A genetic algorithm bred new plans by selection, mutation, and mixing. Two test scenes were used: a simple password game and a mock illegal pet trade. As dialogue rounds grew, talks ran longer without blocks. Accuracy of shared information also rose. A study with forty people backed the realism of the dialogues. Tests without the genetic step did worse, showing its value for long-term adaptation.

Catalog record

CREATORJinyu Cai, Yusei Ishimizu, Mingyue Zhang, Munan Li, Jialong Li, Kenji Tei
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE58 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-19193

Browse the full library · About the Tower of Babel