Do large language models know what they are talking about? This article considers whether large language models (LLMs) have actual knowledge and understanding. The author delves into the philosophical concept of knowledge and the different schools of thought surrounding it, including perception-based constructivism and reason-based rationalism. While LLMs can manipulate language and produce coherent responses, they are limited to the information in their training data and lack consciousness and real-world experiences. The danger lies in treating AI-generated information as purely actionable, as it may lead to an abundance of information that no one truly understands.
Comments