NISHIO Hirokazu[English][日本語]

Intersubjective Model of AI-mediated Communication: Augmenting Human-Human Text Chat through LLM-based Adaptive Agent Pair

image

o3-mini-high.iconThis paper proposes a new model, the Intersubjective Model, in which each user engages in conversation in a unique chat environment, mediated by an LLM-equipped AI agent, which extracts and exchanges information to enable flexible interaction that goes beyond traditional direct information transfer.

Basic Structure: 1.

  • Each user is in a separate chat environment, and each environment has an AI agent. The agent extracts important information from the user's statements and sends it to the agent on the other end. Based on the information received, the agent generates a response to the user.

Features: -Features: -Features

  • Internal feedback: Immediate response between the user and his/her agent, maintaining the rhythm of the conversation
  • External feedback: Agents exchange information with each other to complement the meaning and intent of the overall conversation

Applications: Lively conversations, focused discussions, overcoming language barriers and generation gaps, etc. Example implementation:.

  • Combining a simple chat UI and GPT-4o-based agent functions (extraction and conversation generation), we have built a system that achieves a natural dialogue rhythm.

This model allows users to communicate in their own subjective environment, but with rich communication through AI-mediated information exchange.


This page is auto-translated from [/nishio/Intersubjective Model of AI-mediated Communication: Augmenting Human-Human Text Chat through LLM-based Adaptive Agent Pair](https://scrapbox.io/nishio/Intersubjective Model of AI-mediated Communication: Augmenting Human-Human Text Chat through LLM-based Adaptive Agent Pair) using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.


(C)NISHIO Hirokazu / Converted from Markdown (en)
Source: [GitHub] / [Scrapbox]