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

TL;DR

The paper introduces the Intersubjective Model, an AI-mediated communication framework in which two participants operate within independent environments and communicate through LLM-based agents, enabling real-time message modulation and shared understanding without a common objective space. It presents a prototype text-chat system implementing Extraction and Conversation agents and demonstrates use cases such as lively conversations, focus on topics, language translation, generational bridging, and calm discussions. The authors ground the model in traditional transmission and feedback theories, reframing modulation as an intentional design choice rather than error, and describe a two-loop feedback mechanism that accelerates inner exchanges while preserving outer meaning. They discuss design-space implications, potential benefits, and challenges, and outline a program of empirical evaluation, ethical considerations, and multi-modal/topical expansions toward richer, intersubjective communication in future work.

Abstract

The growing prevalence of Large Language Models (LLMs) is reshaping online text-based communication; a transformation that is extensively studied as AI-mediated communication. However, much of the existing research remains bound by traditional communication models, where messages are created and transmitted directly between humans despite LLMs being able to play a more active role in transforming messages. In this work, we propose the Intersubjective Model of AI-mediated Communication, an alternative communication model that leverages LLM-based adaptive agents to augment human-human communication. Unlike traditional communication models that focus on the accurate transmission of information, the Intersubjective Model allows for communication to be designed in an adaptive and customizable way to create alternative interactions by dynamically shaping messages in real time and facilitating shared understanding between the human participants. In this paper, we have developed a prototype text chat system based on the Intersubjective Model to describe the potential of this model, as well as the design space it affords.

Figures (10)

• Figure 1: Conceptual diagram of communication under traditional model (left) and our Intersubjective Model (right).
• Figure 2: Process of a message sent by Alice transmitted to Bob, mediated by agents. (1) Alice says, "Yesterday I went to a cinema." (2) Agent A in Alice's environment (i.e., proxy of Bob) extracts the information "Alice told Bob that she went to a cinema on Sep. 12" and transmits it to Agent B in Bob's environment (i.e., proxy of Alice). (3) Agent B adds this shared information to its knowledge base. (4) Later, Agent B says, "Oh, I went to a movie theater yesterday!" and Bob receives this message.
• Figure 3: Example 1 - Lively Conversation
• Figure 4: Example 4 - Bridging Generational Gaps
• Figure 5: Example 5 - Facilitating Calm Discussion