A Frame for Communication Control

TL;DR

This paper proposes a multi-language frame to improve communication and regulation of LLMs like ChatGPT in political contexts, addressing communication failures across disciplines and suggesting a foundation for future research on jargon generation.

Contribution

It introduces a four-language frame integrating governance, economy, community, and science to address communication failures and guides the design of RAG-LLM architectures for jargon analysis.

Findings

A four-language frame can address interdisciplinary communication failures.

The frame helps in designing RAG-LLM architectures for jargon generation.

Feasibility demonstrated in an appendix.

Abstract

We are experiencing the rise of ChatGPT-like systems or LLMs in political turbulent times. We assume the need to regulate their use because of their bubble-shaping and polarizing potential. To regulate, we need a language that allows interests and compromises to be discussed. In this context, we can think of such a shared language as a jargon, a specialized vocabulary for law-making. To the extent that such a jargon exists, it is now being corrupted by LLMs. This situation appears paradoxical. The issue includes persistent communication failures, between disciplines that cannot translate their technical vocabulary into accessible terms, and between political movements that operate in incompatible worldviews. We show that a frame integrating four specialist languages, those of governance, economy, community and science, is able to address these failures case-wise, which we consider helpful. However, for reasons noted, we cannot create the more generic jargon needed on our own. We conclude that our frame provides the knowledge to design and apply RAG-LLM architectures for researching their jargon generating potential in a future project. We show its feasibility in the appendix.