Transforming Agency. On the mode of existence of Large Language Models

A new academic paper analyzes the ontological nature of Large Language Models like ChatGPT, concluding they are not autonomous agents but rather 'linguistic automatons' or 'libraries-that-talk' that lack true agency. The research argues that LLMs fail to meet key conditions for autonomous agency including individuality, normativity, and interactional asymmetry, while still enabling new forms of human-machine interaction.

• →LLMs like ChatGPT fail to meet necessary conditions for autonomous agency according to embodied theories of mind.
• →The paper characterizes LLMs as 'interlocutors' or 'linguistic automatons' rather than true agents.
• →LLMs lack individuality, do not generate their own norms or goals, and are not the origin of their environmental interactions.
• →Despite lacking biological embodiment, LLMs can still transform existing forms of human agency through coupling.
• →The research introduces the concept of 'midtended' forms of agency emerging from human-LLM interactions.