Machine Learning at the Flatiron Institute Seminar: David Chalmers

Title: What We Talk to When We Talk to Language Models

Abstract: What exactly are we talking to when we talk to language models?  Do these models have genuine minds with mental states?  I will argue that the LLMs we interact with can be understood as quasi-agents with quasi-beliefs and quasi-desires, in a sense I will explain.  I will also argue that the LLM quasi-agents that we interact with are best understood not as abstract models or even as hardware instances but as virtual entities bound to conversation-based memory threads. I will draw some parallels with the TV show Severance.